__

 

LIBRARY
Michigan State
University

“

I ~

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES Mum on air before one due.

DATE DUE DATE DUE DATE DUE

JUN 0 2 iQQfl u

j  7!
4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

41
~ 7
fit‘i 1

MSU it An Afflnnotivo Action/Equal Opportunity institution
emu.» t

 

 

 

 

 

 

 

 

 

 

 

 

ESTIMATION OF THE GENETIC STRUCTURE OF AN ELITE
SOYBEAN POPULATION AND IT’S APPLICATION TO A
SOYBEAN BREEDING PROGRAM

By
Clay Hurd Sneller

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Crop and Soil Science

1991

ABSTRACT

ESTIMATION OF THE GENETIC STRUCTURE OF AN ELITE SOYBEAN
POPULATION AND IT'S APPLICATION TO A SOYBEAN BREEDING
PROGRAM

ray

Clay Hurd Snellcr

The source and frequency of genes, and the genetic distance
between lines. was estimated with coefficient of parentage (CP) and a
similarity index (SI) based on marker loci for a set of 62 elite soybean
lines. The genetic base of this population derived from 27 ancestral
parents with "Lincoln" and "Mandarin Ottawa" contributing 30.6% and
17.7% of the parentage respectively. The average 81 value between
the lines was 0.64 and the average CP was 0.26. While the average
relationships suggested a narrow genetic base, pairs of lines could be
identified that theoretically shared few genes and certain lines were
identified as outliers. There was a poor correlation between the SI
and CP values. Selection of lines based on maturity and yield criteria
appeared to result in a more inbred population than a general set of
public cultivars with a broader maturity range. The restriction in the
elite population appeared to be due to selection for similar maturity
versus selection for similar yield potential in a narrowly defined
environment. __

Genetically similar and distance elite lines were crossed and
approximately 55 F2z3 or F4:5 families were derived from each cross
and field tested for seed yield. plant height. and date of maturity. The
generalized variance. and the genetic variance for individual traits
were estimated for each population. A third parental distance (PCD)
was derived from a principal component analysis of CP and SI data.
The variance parameters of a population increased as the parental

genetic distance increased. regardless of the method of estimating the
distance. There was a stronger association of the measures with the
generalized variances than with the genetic variances. No measure
was significantly associated with yield genetic variance while all
seemed predictive of the genetic variances for maturity and height.
The utility of the distance measures appeared to be limited to
identifying which populations would have an increased likelihood of
having an above average genetic variance for the individual traits. The
PCD appeared to be the best predictor of the variance parameters.

Table of Contents

Literature Review 1
1. Genetic Distance in Plant Breeding 1

II. Estimating Genetic Distance 4

A. Quantitative Data 4

B. Qualitative Data 8

C. Pedigree Data 1 2

111. Application of Genetic Distances to Soybean Breeding 1 5

List of References 1 7

Estimation of the Genetic Structure of an Elite Soybean Population 2 1

Introduction 2 1
Materials and Methods 23
Coefficient of Parentage Data 23
Marker Loci Data 2 7
Results 29
Coemcient of Parentage Data 29
Marker Loci Data 38
Discussion 45
List of References 52

Appendices 54

iv

(Application of Genetic Distances to a Soybean Breeding Program
Introduction
Materials and Methods
Results
Discussion

List of References

70
7O
73
77
90
94

ListofTables

Estimation ofthe Genetic Structure ofan Elite Soybean Population

Table 1. Summary of the source and name of the 62 elite lines, the
average, minimum, maximum, standard deviation, and correlation of
the coefficient of parentage (CP) and similarity index (SI) values for
each line as compared to the other 61 lines.

24-25

Table 2. Summary soybean cultivars included in the general
population.
26

Table 3. The average coefficient of parentage (CP) between lines of
the general population and it's maturity subdivisions and between the
elite lines with or without a subset of the elite lines that appeared to
be outliers.

30

Table 4. Summary of significant correlations between the
contributions from ancestral parents (APs) to the elite population (re)
and the broader population (rooqv), their contributions to the elite
population (APC) along with the primary progeny line derived from the
correlated ancestral parents.

33

Table 5. Correlations of the contributions from ancestral parents that
were a source of at least 1% of the genes in the elite population with
the principal component scores from the analysis of this data set and
the percentage of the variation in this data set that is accounted for by
the first three principal components.

37

Table 6. Ancestral parent contributions to the five clusters of elite
lines obtained by Ward's clustering method.
40

Table 7. Summary of allele frequencies and heterogeneity in the elite
population for nineteen marker loci and the correlation of the
frequencies with the first three principal components.

41

vi

Vii

Application of Genetic Distance. to a Soybean Breeding Promm

Table 1. Summary of the parentage. coefficient of parentage (CP),
similarity index (81). and the principal component distance (PCD)
between the parents for all crosses used to generate the segregating
populations along with the number of families in each population and
the years tested.

75

Table 2. Summary of the generalized variances (GV) and the maturity,
height. and seed yield progeny variances (PV) estimated in 1989.
1990 and 1989-90 for each population.

78

Table 3. Spearman's rank correlations of the generalized variances
(GV) and the maturity, height. and seed yield progeny variances (PV)
of the segregating populations with the coefficient of parentage (CP).
similarity index (SI). and the principal component distance (PCD)
values of the parents of the populations. summarized by year(s). 9
7

Table 4. Summary of the R2 values and their significance from the
regressions of the GV and maturity. height and seed yield progeny
variances (PV) of the segregating populations on the coefficient of
parentage (CP), similarity (SI) and principal component distance
(PCD) values of the parents of the populations. summarized by year(s).
80

Table 5. The average seed yield progeny variances of the populations
whose parents had coefficient of parentage (CP). similarity index (SI)
or principal component distance (PCD) values that were either higher
(High) or lower (Low) than the mean of these measures, summarized

byyears.
90

ListofFigures

EetimationoftheGeneticStructureofanEliteSoybeanPopulation

Figure 1. The percentage of the elite and general populations's
parentage that derived from the ancestral parents that contributed at
lest l .0% of the parentage.

32

Figure 2. The percentage of the elite population's parentage that was
derived from the ancestral parents that contributed at least 1.0% of
the population's parentage.

36

Figure 3. Plot of the first three principal components (PCAl, PCA2,
and PCA3) from the principal component analysis of the ancestral
parent contribution data of the 62 elite lines. Shapes correspond to
the five clusters derived from hierarchical clustering of the elite lines
using the same data and Ward's clustering criterion (see Figure 4).

33

 

Figure 4. Results of clustering the 62 elite lines using the ancestral
parent contribution data set after pooling the highly correlated
contributions and using Ward's minimum variance clustering criterion.

39

Figure 5. Plot of the first three principal coordinates (PCOI. PC02,
and PCO3) from the principal coordinate analysis of marker loci data
of the 62 elite lines. Shapes correspond to the five clusters derived
from hierarchical clustering of the elite lines using the same data and
the average clustering criterion (see Figure 6).

43

Figure 6. Results of clustering the 62 elite lines using the marker loci
data set and the average clustering criterion.
44

viii

ix

Figure 7. Plot of the first three principal components (PCAl, PCA2.
and PCA3) from the principal component analysis of the ancestral
parent contribution data of the 62 elite lines. Shapes correspond to
the five clusters derived from hierarchical clustering of the elite lines
using the marker loci set and the average clustering criterion (see
Figure 6).

 

46

Figure 8. Plot of the first three principal coordinates (P001, P002.
and P003) from the principal coordinate analysis of marker loci data
of the 62 elite lines. Shapes correspond to the five clusters derived
from hierarchical clustering of the elite lines using the ancestral
parent data set and Ward's clustering criterion (see Figure 4).

47

Application of Genetic Distances to a Soybean Breeding Program

Figure 1. Results of the linear regression of the 1989 estimates of the
generalized variances (GV) of the segregating populations on the
coefficient of parentage (CP) of the parents of the

8 1

Figure 2. Results of the linear regression of the 1990 estimates of the
generalized variances (GV) of the segregating populations on the
similarity index (SI) of the parents of the populations.

82

Figure 3. Results of the linear regression of the 1989-90 estimates of
the generalized variances (GV) of the segregating populations on the
principal component distance (PCD) between the parents of the
populations.

83

Figure 4. Results of the linear regression of the maturity genetic
variances of the segregating populations on the principal component
distance (PCD) between the parents of the populations.

85

Figure 5. Results of the linear regression of the height genetic
variances of the segregating populations on the principal component
distance (PCD) between the parents of the populations.

87

Figure 6. Results of the linear regression of the seed yield genetic
variances of the segregating populations on the principal component
distance (PCD) between the parents of the populations.

88

X

7. Results of the linear regression of the 1989 estimates of the
seed yield progeny variances of the segregating populations on the
principal component distance (PCD) between the parents of the
populations.

89

LiteratureReview

I. GeneticDistanceandPlantBreedlng

The existence of genetic diversity is the cornerstone of
improving plants to meet human needs. Our present crop species
were selected from among the vast array of species in the plant
kingdom and for centuries farmers selected among existing diverse
genotypes to improve these species. Later plant breeders began
crossing different genotypes to create further diversity by recombining
the genes of one parent with the difi'erent genes of the other parent.
The success of crossing depends on the assumption that the two
parents possess different genes (i.e. that there is genetic distance
between them). Furthermore. breeders make crosses with the
purpose of generating new and improved genotypes for a particular
trait and therefore require that the crossed parents are genetically
distant in regards to trait to be improved.

The necessity of genetic distance between parents to obtain
breeding progress can be seen in the gain from selection equation.

AG = S h2 = S 02a/02p = S (0251/02a + 62d + 02m + 02¢ + 0%)

AG = gain from one cycle of mass selection
S = selection difl'erential
h2 = narrow sense heritability
0% = additive genetic variance = 2 (2p1q1a12)
a1 = average effect of an allele substitution, [31+dj(Q1-p1)]
pi. m = frequency of the two alleles at the ithl locus
a; = absolute deviation of the homozygous genotypes at the
ith locus fromtheir mean
d1 = deviation of the 1th heterozygote from the mean of the
two homozygous genotypes
091, = phenotypic variance
02m = non-additive genetic variance
62¢ = environmental variance

628: = genotype x environment variance

Gain from selection should increase as additive genetic variance
increases assuming that the population mean and phenotypic variance
remain constant. In a segregating population that is generated from
crossing two individuals, additive genetic variance at the 1th locus is

l

2

maximized when p) = q; = 0.5. a situation that occurs when each

parent possess different alleles. Additive genetic variance is the sum
of the additive effects from all loci that affect the measured trait and
will increase as the genetic distance between two parents increases as
this should increase the number of segregating loci that affect the trait
increases in the population.

The objective of plant breeding for metric traits is to select a
genotype or a population of individuals that has a more desirable value
for the selected trait than the parents. The mean of the progeny (u')
derived from selected individuals can be expressed in terms from the
gain from selection equation as

AG = u'-u = Sh2 = (us,-u)h2
u' = (us—u)h2+u

u = mean of the parental population
us = mean of the selected individuals

 

The heritable portion of the selection difi'erential is added to the mean
of the parental population to produce u'. The ideal situation for
obtaining progeny with a high mean is when the parents have a high
mean (elite parents) and are genetically diverse. A high parental mean
can be assured by thorough testing of the parents in the desired
environment but aprlori knowledge of the genetic diversity between
the parents can only be estimated or inferred.

The amount of genetic variability in a population is one
parameter that can be measured in a hybrid population that is
influenced by the genetic distance of the parents (Cowen and Frey.
1987a). A second parameter is the amount of mid-parent heterosis
(Hm) exhibited by the hybrid population (Falconer, 1981):

Hm = .2 (113’:2
Y: = P - P'
p = frequency of one of the two alleles at the 1th of n loci in
parent one
p' = frequency of the same allele in parent two

3

It is apparent from this equation that heterosis will increase in
proportion to the degree of genetic divergence of the parents (y)
when a directional dominance exists over all loci (i.e. 2d) = 0). The
genetic variance and heterosis exhibited in a hybrid population are
functions of genetic efi'ects (a1 and d) respectively) as well as parental
genetic distance such that the amount of either may not always be
proportional to parental genetic distance.

A third parameter associated parental genetic distance is the
number of trangressive segregants in a hybrid population. Assuming
additive or dominant gene action for a given trait. then a high
transgressive segregant has accumulated more alleles for increased
expression of the trait than the high parent (or decreasing alleles than
the low parent). The frequency of such progeny from parents with
equal phenotypes should increase as parents become more diverse at
loci that affect the trait. The frequency of transgressive segregants is
also dependent on the distribution of the increasing and decreasing
alleles between the parents: a parent with all decreasing alleles is
quite distant from a parent with all increasing alleles but a cross of
these parents would not produce any transgressive segregants. Thus
these three parameters measured of hybrid populations that are
afiected by parental genetic distance are also afi‘ected by other factors
and may not be directly proportional to the genetic distance of the
parents.

The accuracy of the estimate of genetic distance between the
parents will vary with the estimation method. One approach is to
compare the parents on the basis of various heritable traits and
attempts to measure genetic differences through phenotypes. This
approach has been used with both quantitative traits such as
morphological traits and with qualitative traits such as polymorphisms
at isozyme and restriction fragment length polymorphism (RFLP) loci.
Another approach is based on the examination of pedigree data under
a set of assumptions concerning ancestral genealogy, ancestral
genotypes and the genetic transmission between parents and
offspring. These approaches. their assumptions, biases. and their
applicability to a breeding situation are discussed below by the type of
data that is used to determine the genetic distance. While there is a

4

large body of research where inferred genetic distances are related to
breeding behavior or where quantified genetic distances are used only
to infer relationships among genotypes, this author has only attempted
to review methods of estimating genetic distance and research where
quantified genetic distances between parents were related to some
form of breeding behavior in the hybrid population. The terms
genotype and population are used synominously

ILEctimatingGeneticDistance
A. Quantitative”

Morphological data can be used to estimate the relationship
between two populations by obtaining a reliable estimate of the
phenotype of the populations for one (univariate) or more
(multivariate) traits and statistically comparing them. This approach
actually results in a statistical distance between the populations and
inferences of a genetic distance are made under the assumption that
that phenotypic difi‘erences reveal underlying genotypic difl'erence. A
large phenotypic difference will result in a large distance regardless of
the extent of underlying genetic differences. Accurate inferences are
more likely to be made when highly heritable traits are used and as
more traits are examined since this will assay for genetic differences
at more loci.

There are several statistical methods for estimating the genetic
distance between two populations with phenotypic data. One of the
simplest approaches is to calculate the Euclidian distance (ED)
between the populations (Goodman, 1972) where the distance
between the 1th and 101 population is

any = ii (xii. - 19m 11/2
"I
Xm, Xjk = mean of the l:th of n traits for the ith and jth populations

respectively that have been standardized by dividing by
the standard deviation of kth trait.

5

Each population can be visualized as a point in n dimensions
corresponding to the n traits where the mean for the kth trait for the
population is the coordinate for that point along the kth axis. ED” is
the distance between the ith and jth point in this multidimensional
space. ED” is also equivalent to the distance between the principal
components derived from an analysis of the correlation matrix
between the traits (Goodman, 1972) where

Eng = I “£30m. 400211”

Ylk, ij = coordinates of the kth of n principal components for
the 1th and 1th populations respectively.

In a principal component analysis the original n axes that
corresponded to each trait are rotated in space to correspond to n
new variates that are uncorrelated to each other. The population
points remain in the same position as before but their locations in the
n dimensional space are now defined by coordinates for the new axes.
The new coordinates for the 1th population are it's principal
component scores.

The Euclidian distances such as Pearson's (1926) coefficient of
racial likeness or Sokal's (1961) distance can be used when all the
measured traits are uncorrelated (Goodman. 1972). This would seem
to be an unlikely scenario especially when many traits are compared
and ignoring the correlations could result in an exaggerated genetic
distance when the correlations have a genetic basis and would be
similar to giving some gene differences more weight than others in
the distance measure.

Mahalanobis (1936) first addressed the problem of
intercorrelated traits by calculating a generalized distance (D) where
the Euclidian distance is adjusted by the common within-population
correlation matrix such that the distance between the 1th and jth
population is

Dr = Im-xjrn-l (xi-xplm

Di) = [firm-WW mil/2

6

X), x, = vectors of the standardized means of the n traits for the
1th and jth populations
R'1 = inverse of the correlation matrix of the n traits
1k = eigenvalue of the kth principal component

D is the multivariate generalization of "Student's" t test (Hotteling.
1954) which is commonly used to test the equality of two means and
D2 can be used to calculate Hotteling‘s ’1‘2 test of the similarity of two
mean vectors in conjunction with a multivariate analysis of variance.
All eigenvalues equal unity when all traits are uncorrelated and D then
equals ED. A problem with D is that principal components with very
small eigenvalues and which therefore account for very little of the
overall variability and may not be biologically significant, can have an
inflated contribution to D, negating the contribution from more
important principal components (Goodman, 1972).

An alternative is to include only some of the standardized
principal components in the distance (Goodman, 1972). This
approach projects the population point in a space with fewer but
hopefully more biologically significant dimensions. Goodman (1972)
suggested using the kth principal component in the distance estimate
only when 1}; >= k. A less conservative approach is to use all principal
components where 3.}; >= 1.

Calculating and interpreting these distances becomes more
complex when there is heterogeneity between the covariance matrices
of the populations being compared as this can result in D1] differing
from D11 (Atchley et aL, 1982). The heterogeneity of covariance
matrices reflects different relationships of the traits within different
populations and may be due to biological differences brought about by
genetic and developmental differences and the interaction of these
factors with the environments where the traits were measured
(Atchley et at. 1988) . The interaction of the populations and the traits
with the environment requires that all populations and traits be tested
simultaneously in multiple environments and new populations could
not be added to this data base and compared to the others without
remeasuring the old populations.

7

When a breeder tries to improve a particular trait by crossing
parents with high means for that trait then the parents will not show
much phenotypic variability for that trait. In this situation the breeder
would have to measure phenotypic variability in other morphological
traits and hope that it is predictive of genetic variability for the
desired trait. Narrow breeding populations may lack sufficient
variability for other morphological traits for successful application of
this approach especially if the other traits are correlated to the trait
that is being improved. The use of quantitative traits has the potential
to assay genetic distance at many loci as there are perhaps hundreds
of gene differences between two populations that reside at the
extreme ends of the expression range with multiple morphological
traits. though it seems unlikely that this extreme range would exist in
an elite breeding population.

Genetic distances are pairwise comparisons between populations
but they can also be used to elucidate a broader relationship that may
exist among all genotypes by coupling the resulting distance matrix
with a clustering technique. This is a common practice in taxonomy
and classification work that also can be applied to a breeder's parental
populations.

There are numerous reports of the use of quantitative trait
distance techniques in taxonomy. classification. and evolution studies
but there are relatively few studies that have related these parental
distances to any form of breeding behavior exhibited by a hybrid
population. The morphological distance between parents has been
positively correlated to the mid-parent heterosis for grain yield
exhibited by hybrids in rapeseed (Brassica napus L.)(Lefort-Buson et
al.. 1987). soybeans (Glycine max L. Merr.)(Chauhan and Singh. 1982).
wheat (Triticum aestivum L.) (Shamsuddin, 1985; Cox and Murphy.
1990). dry beans (Phaseolus vulgaris L.) (Ghaderi et al.. 1984). peanut
(Arachis hypogaea L.)(Arunachalam. 1984: lsleib and Wynne. 1983).
maize (Zea mays L.) (Prasad and Singh. 1986) and tomato
(Lycopersicum esculentum Mill.)(Maluf et al.. 1983) while no
relationship was found in faba bean (Vicia faba L.) (Ghaderi et al.. 1984)
or cats (Avena sativa L.) (Cowen and Frey. 1987a). Cox and Murphy
(1990) found that the value of their morphological data in predicting

8

heterosis in F2 wheat populations depended on the environment in
which the data was collected.

There are few studies relating distances based on morphological
data to genetic variance and the number of transgressive segregants in
segregating populations despite the importance of these parameters to
the plant breeder. The Euclidian distance between oat parents based
on morphological data was negatively correlated with the number of
transgressive segregants and the generalized variance (Cowen and
Frey. 1987a) which is a measure of the overall genetic variance in a
population comprised of the variation of all measured traits (Sokal.
1965: Goodman. 1968).

BQmfltativeData

Actual genetic distances between parents can be estimated by
using allele frequency data from simply inherited polymorphisms. In
plants. suitable data can be derived from isozyme. RFLP. disease
resistance. and some qualitative morphological trait loci. hereafter
referred to collectively as marker loci (ML). This data has an
advantage over quantitative morphological variation which can only
infer genetic difference at an unknown number of loci as it directly
assays for differences at known loci and provides a quantified estimate
of actual genetic differences. The ML phenotypes are generally not
influenced by the environment, thus eliminating a major limitation of
morphological data and allowing new populations to be easily added to
an existing data base of ML profiles.

The general approach to calculating a genetic distance from ML
data is to determine the frequency of all ML alleles in the parental
populations and comparing these ML profiles using a distance
equation. An early statistic for measuring the distance between
populations calculated the correlation (Fst) between random gametes
within the two populations relative to the correlation of the gametes if
the populations were combined. This was first proposed by Wright
(1965) and modified by Net (1965) for multiple alleles:

9
Fst = (- O'Pij) / (Pij)

covariance of the frequencies of the jth and k1:11 alleles

mean frequencies of the jth and kt-h alleles for the two
populations

0'13ij
Pjv Pk

A drawback to this statistic is that Fst can be negative.

Sokal and Sneath (1963) presented a formula for calculating the
probability that a randomly sampled allele from one population is
identical by state (IBS) to a randomly sampled allele from the same
locus in the other population. averaged over all sampled loci (13):

Is = (l/ningPrjaPrjb

pija. pub = the frequency of the 1th of m alleles at the jth of n loci in
populations A and B respectively

Is measures the probability of gametic similarity between populations
and ranges from zero (no alleles in common) to unity when each
population is fixed for the same alleles at all loci. Two populations
which have equal p11 values but where 0<pg<l will not produce Is
values of one and some populations with identical gene frequencies
may appear more distant from one another by this measure than two
populations which have different gene frequencies (Spiess. 1977).
Nei's (1972) measure of the genetic identity (IN) and distance (Du) are
commonly used to compare populations and will result in a identity
score of unity (or a distance score of zero) when two populations have
identical allele frequencies regardsless of the value of pg:

IN = 1s / l { égpgazifgépgfi) } / n211/2
DN = -log IN

From a breeding stand point it is important to note that loci where
pga=pgb and 0<pga<l will still contribute to the heterozygosity of a
hybrid between the two populations despite producing an IN value of
unity. If the parental populations are randomly mating then the
heterozygosity of the hybrid population will be equal to that of it's

10

parents and the IN value of one accurately indicates that the hybrid
population will be the same as the parents. When the the parental
populations are inbred. an IN value of unity for such loci will not reflect
the heterozygosity that will occur in the hybrid population in
proportion to allelic heterogeneity at the jth locus. Upon inbreeding.
the hybrid population will return to the genotype frequency at the jth
locus of the parents but there will have been an opportunity for
recombination to produce new genetic variation when IN at other loci
is less than unity or when other loci with an unity IN are
heterogeneous. Sokal and Sneath's ls measure of distance is
predictive of hybrid heterozygosity and therefore would seem to be
more relevant to plant breeders.

Hedrick (1970) proposed a measure called the probability of
genotypic identity (Iij) which is based on genotype frequencies
instead of allele frequencies.

IHlk = (épupik) / [1/2( £11ij + .i': Pikzll

Prj. pa; = frequencies of the 1th of n genotypes in the jth and kth
populations

This statistic has certain appeal in evolutionary studies but seems less
applicable to breeding situations where gamete frequency and thus
allele frequency is a primary interest.

Principal component analysis of the correlation matrix of allele
frequencies in conjunction with Goodman's approach can be used to
produce a distance measure where the variables are uncorrelated. As
with quantitative data. all of these distance measures are pairwise
comparisons between two populations but the resulting distance
matrices can be coupled with a clustering technique to ascertain an
overall structure between all populations.

As with quantitative data. the effectiveness of this approach
depends on sampling and the accuracy of estimates of genetic distance
will increase when genomic diversity is assayed with more ML. The
development of isozyme and RFLP technology has made this approach

ll

feasible in species where few polymorphic ML were previously
available.

ML are generally regarded to be neutral in their efl'ect on other
traits and difi'erences ML must accurately predict differences at the
loci that affect the trait being improved to be of value to a breeder.
The ability to predict heterogeneity at loci controling the object trait
depends on the number and chromosome distribution of the loci
controlling the trait. the number and distribution of the ML. and the
degree of linkage disequilibrium between the two sets of loci. If the
assayed ML are randomly sampled then the ML differences between
populations should reflect their overall genomic diversity. The
calculations of Chakraborty ( 198 1) indicate that an unrealistically large
number must be assayed to obtain an accurate estimate of genetic
distance when genomes contain thousands of independent genes but
with the organization of genes into linkage groups located on
chromosomes. one ML may be able to assay for differences between
other genes within it's linkage group. The use of RFLPs and saturated
genetic maps it is now possible to use ML to assay for genetic diversity
across virtually all chromosome segments. ML difi'erences will more
accurately assay differences at other loci when a linkage disequilibrium
exists between the loci: a situation that is more likely to occur in
autogamous species and in populations that have a narrow genetic base
that further increases inbreeding.

There is accumulating evidence (Frei et al.. 1986: Graef et al..
1989: Kahler and Wehrhahn. 1986: Nienhuis et aL. 1987: Osborn et
al.. 1987. Stuber et al.. 1987) that ML heterogeneity in hybrid
populations is frequently associated with quantitative trait variation.
indicating that a linkage disequilibrium between ML and loci that
affect the a quantitative trait can be considerable. As chromosome
segments are identified that contain genes that affect the breeder's
object trait then ML linked to those segments can be selected and
assayed for polymorphisms. This approach will be more cost efficient
and will result in a distance measure that is unbiased by data
generated by ML not associated with relevant genes.

There have been several attempts, with mixed results. to
correlate the performance of hybrids between inbred maize lines.

12

which is generally dependent on the extent of heterosis. to a ML
derived genetic distance between the parents. Price et al. (1986) and
Lamkey et aL (1987) found a significant but low positive correlation
between hybrid yield and inbred parental distances based on isozymes
indicating that the diversity estimate had little value in predicting
hybrid performance. Other researchers using isozymes (Hunter and
Kannenberg. 1971; Hadjinov et al. 1982) and RFLPs (Godshalk et aL.
1990) have found insignificant correlations between hybrid yield and
parental distances. Frei et al. (1986) found the ML distance between
parents to be predictive of hybrid performance only when there was a
known. pedigree relationship between the parents: a situation that
increases inbreeding and the probability of linkage disequilibrium
between ML and other genes. Lee et al (1989). using RFLPs to
estimate parental diversity within and across chromosomes. found a
significant correlation between parental diversity and hybrid yield and
that the diversity of some chromosomes was more predictive of yield
than other chromosomes. Their analysis of specific combining ability
(SCA) produced similar results while others have reported no
association of SCA and parental diversity based on isozymes (Hunter
and Kannenberg. 1971: Heidrich-Sobrinho and Cordeiro. 1975:
Hadjinov et al.. 1982).

A general conclusion from this work is that ML appeared to have
some utility in assigning inbreds to general heterotic groups but that
they had little value in predicting parameters such as heterosis. GCA.
and SCA which are so important in maize. The work of Lee et aL
(1989) suggests this approach will be more successful when only the
diversity of relevant chromosome segments is used. No research has
been performed relating ML derived distances to the amount of
genetic variance or the number of transgressive segregants in hybrid
populations.

CPedipeeData

The known genealogical relationship between individuals and
populations can be used to infer the probability that they share
common alleles at a locus. Wright (1920) used the complete

l3

pedigrees of two population to calculate a coefficient of probable
relationship which not only described the genetic similarity between
two populations. but also the inbreeding coefficient of their progeny.
It is the probability that a randomly sampled allele from one individual
(or population) is identical by descent (IBD) to an allele from the same
locus from the other individual. Malecot (1948) called this value the
kinship coefiicient. Kempthorne (1969) refered to it as the coefficient
of parentage (CP) while others have called it the coefficient of
coancestry. (l-CP) is the theoretical distance between populations.
The CP between populations A and B can be expressed as

It
rAB = .2“: (1/2)n +13 +1 (14-de

where C). whose inbreeding coefi'icient is M. is a common ancestor of
A and B to the n) and p) degrees respectively and where there are k
different exclusive ways of relating A and B to their common
ancestor(s). Kempthorne (1969) presented a working formula of

rAB = l/2(1'Ax+ rAv)

where rAx and rAy are the CPs between A and the parents of B. X and Y.
A CP value of one indicates the populations are identical while a C? of
zero indicates that the populations theoretically share no IBD genes.
The CP is easy to calculate and encompasses all of the genome thus
eliminating the problem of sampling genes that exists with the other
approaches.

A CP is based on the assumptions that (i) all ancestral parents
are completely unrelated to one another (i.e. no genes in common).
(it) an estimate of the level of inbreeding in each ancestral parent and
(iii) that each parent contributes equally to all offspring. The accuracy
of the estimate depends on the validity of the assumptions. It is quite
unlikely that the first assumption is true for all loci as it dictates that
each ancestral parent posses an unique allele at each locus. The level
of ancestral parent inbreeding can be assured in inbred lines though
gametic heterogeneity can still occur if the ancestral parent is not a
pure line. as is often the case with plant introductions and landraces.

14

The ancestral parent inbreeding can only be inferred for any locus in.
an ancestral parent where cross pollination occurs. It seems certain
that the assumption of equal biparental contributions to all progeny
must be violated to some extent in cultivars developed through many
cycles of intense selection and particularly for loci that have a large
and highly heritable effect on a selected trait. St. Martin (1982)
estimated that in soybeans (2n=40). 88% of the lines derived from a
biparental cross will obtain from 40 to 60% of their genes from one
parent (assumed percentage is 50%) and that a breeder would be
unlikely to select a line with even 70% of it's genes from one parent.
These calculations indicate. at least for genes with a small effect on a
selected trait. that the assumption of equal parental contribution may
not be seriously violated. The validity of the assumptions will vary
from trait to trait and from gene to gene. The violations may cancel
out when averaged over all genes that aflect a trait (unless there is a
directional bias due to selection) suggesting that this approach may
produce more accurate estimates of distance for traits that are
controlled by many genes.

While the coemcient of parentage is usually calculated between
the two populations being compared. one can also calculate the
coefiicient of parentage between each population and all the ancestral
parents that appear in the pedigree of the populations (Souza and
Sorrels. 1989). This method describes each population with a set of
variables that estimate the probability that the population carries an
allele the is IBD to that of an ancestral parent. Multivariate statistic
techniques such as principal components and clustering can then be
applied to this data set.

There has been very little work relating the CP between parents
to the breeding behavior of their hybrid populations. A low. though
significant. positive correlation was reported in rapeseed (Lefort-
Buson et aL. 1987) while the association was not significant in cats
(Cowen and Frey. 1987b) and wheat (Cox and Murphy. 1990). When
predicting the heterosis in F2 wheat populations. Cox and Murphy
(1990) suggested using the CP to select diverse parents after the
parental candidates had already been selected for their performance
and phenotypic divergence. Cowen and Frey (1987b) found a

15

significant association between the CP of the parents of segregating oat
populations and the number of transgressive segregants for plant
height but not for bundle weight. grain yield. straw yield. heading
date. and harvest index.. They did find a significant association
between the CP of the parents and the generalized genetic variance of
the segregating population.

m. ApplicationofGeneticDistancestoSoybeanBreeding

The genetic base of commercial soybeans is known to be quite
narrow. The Committee on Genetic Vulnerability of Major Crops
(1972) analyzed the pedigree of 62 northern cultivars and found that
88% of the parentage derived from only 12 plant introductions (PIs).
Delanney et al. (1983) analyzed the pedigree of 158 US and Canadian
northern cultivars (maturity group 00-IV) released from 1971 to 1980
and found that 50 Pls accounted for all the parentage with 10 of these
accounting for 80% of the parentage. Two Pls. "Mandarin" and
"Mandarin Ottawa" accounted for 30% of the parentage and while
these were assumed to be genetically distinct. there is evidence that
the may genetically similar (Cox et al.. 1985: Kiem et aL. 1989). From
1951 to 1980 the was an overall reinforcement of the predominance
of certain major P13 in the adapted gene pool and little change in the
constitution of the genetic base. The gene pool is even more
restricted in southern adapted lines where 82% of the germplasm
released from 1971-1980 can be traced to seven Pls. St. Martin
(1982) calculated an average CP of 0.25 (i.e. equivalent to that
expected among inbred half-sibs) among 25 cultivars released from
1976 to 1980 under the assumption that ”Mandarin " and "Mandarin
Ottawa” are identical while Cox et al. (1985) reported an average CP of
0.19 among 39 cultivar released from 1971 to 1981. The CPs
represent the minimum distance between lines assuming that the Pls
are completely unrelated. an assumption that may not be entirely valid
in soybeans (Cox et al.. 1985: Delanney. 1983; Kiem et al.. 1989). Cox
et al. ( 1985) also calculated a similarity index value between the
cultivars in their study using qualitative data from isozyme and

16

morphological loci and reported an average similarity index value of
0.63 for lines released from 1971-1981.

Ultimately the genetic base of commercial soybeans must be
broadened by utilizing new PIs to continue to achieve breeding
progress. At present only 50 of 8000 PIs have been extensively used
and there are some PIs with acceptable agronomic merit. The initial
use of a PI in a breeding program generally results in progeny that are
unacceptable and further breeding is necessary to develop a PI-
derived progeny that is agronomically desirable. In the meantime the
breeder must continue to recombine the genes in the adapted gene
pool to produce better cultivars. The narrow genetic base presents a
problem as it is apparent from the reviewed research that a cross
between two randomly selected adapted lines will result in
considerable inbreeding in the resulting F1 and this would reduce the
number of segregating loci in subsequent selfed populations. It would
be very useful to have a simple and accurate method of estimating the
genetic distance between adapted soybean lines that would allow a
breeder to maximize the use of the genetic diversity that does exist in
this narrow gene pool.

Coefficient of parentage can be easily applied to adapted
soybeans as there is extensive pedigree data available on most high
yielding adapted lines. The major shortcoming with this technique in
soybeans is the high probability that some PIs are related. a violation of
one of the CP assumptions. A distance measure based on qualitative
data could also be used with soybeans as the adapted soybeans exhibit a
moderate amount of variability at a number of morphological. isozyme.
and RFLP loci (German. 1983; Doong and Kiang. 1987; Kiem et al..
1989). The narrow genetic base and self pollinated nature of soybeans
would increase the linkage disequilibrium between ML and associated
quantitative trait loci such that dissimilarity at ML will be more likely
to indicate dissimilarity at these linked loci than in a broad based
random mating population.

List ofReferences

list of References

Arunachalam. V.. A. Bandyppadhyay. S.N. Nigam. and RW. Gibbons. 1984. Heterosis in
relation to genetic divergence and specific combining ability in groundnut (Arachis
hypogaeaL). Euphytica 33:33-39.

Atchley. W.R.. E.V. Nordheim. F.C. Gunsett. and P.L. Crump. 1982. Geometric and
probabilistic aspects of statistical distance functions. Syst. 2001. 31:445-460.

Atchley. W.R.. S. Newman. and DE. Cowley. 1988. Genetic divergence in mandible form
in relation to molecular divergence in inbred mouse strains. Genetics: 120:239-253.

Chakraborty. R. 1981. The distribution of the number of heterozygous loci in an
individual in natural populations. Genet. 98:461-466.

Chauhan. V.S.. and BB. Singh. 1982. Heterosis and genetic variability in relation to
genetic divergence in soybean. Indian J. Genet. 42:324-328.

Cowen. NM. and K.J. Frey. 1987a. Relationships between three measures of genetic
distance and breeding behavior in cats (Arena sativa L.). Genome 29:97-106.

Cowen. N.M.. and ILJ. hey. 1987b. Relationship between genealogical distance and
breeding behavior in oats. Euphytica 36:413-424.

Cox. T.S.. Y.'l‘. Kiang. M.B. Gorman. and D.M. Rodgers. 1985. Relationship between
coefficient of parentage and genetic strnilarity indiws in the soybean. Crop Sci. 25:529-
532.

Cox. T.S.. and J.P. Murphy. 1990. The effect of parental divergence on F3 heterosis in
winter wheat crosses Theor. Appl. Genet. 79:241-250.

Delanney. X.. D.M. Rodgers. and RG. Pahner. 1983. Relative contribution among
ancestral lines to North American soybean cultivars. Crop Sci. 23:944-949.

Doong. J.Y.H.. and Y.T. Kiang. 1987. Cultivar identification by isozyme analysis. Soy.
Genet. Newsl. 14:189-225.

Falconer. D.S.. 1981. Introduction to quantitative genetics. Longman House. New York.

Frei. O.M.. C.W. Stuber. and MM. Goodman. 1986. Use of allozymes as genetic markers
for predicting performance in maize single cross hybrids. Crop Sci. 26:37-42.

Ghaderi. A.. M.W. Adams. and A.M. Nassib. 1984. Relationship between genetic
distance and heterosis for yield and morphological traits in dry bean and faba bean.
Crop Sci. 24:37-42.

Godshalk. 15.8.. M. Lee. and KR. Lamkey. 1990. Relationship of restriction fragment

length polymorphisms to single-cross hybrid performance of maize. Theor. Appl.
Genet. 80:273-280.

17

18

Goodman. M.M.. 1968. A measure of ”overall variability" in populations. Biometrics
24:189-192.

Goodman. M.M.. 1972. Distance analysis in biolog. Syst. 2001. 21:174-186.

Gorman. MB. 1983. An electrophoretic analysis of the genetic variationin the wild and
cultivated soybean germplasm. Ph.D. dissertation. Univ. of New Hampshire.

Gorman. M.M.. Y.T. Kiang. Y.C. Chiang. and RC. Palmer. 1983. Electrophoretic
classification of the early maturity groups of named soybean cultivars. Soybean Genet.
Newsl. 9:143-156.

Graef. G.L.. W.R. Fehr. and S.R. Cianzio. 1989. Relation of isozyme genotypes to
quantitative characters in soybeans. Crop Sci. 29:683-688.

Hadjinov. M.I.. V.S. Scherbak. N.I. Benko. V.P. Gusev. T.B. Sukhorzhevskaya. and L.P.
Voronova. 1982. Interrelationship between isozymic diversity and combining ability
in maize lines. Maydica 27: 135-149.

Hedrick. P.W.. 1971. A new approach to measuring genetic similarity. Evol. 25:276-
280.

Heidrich-Sobrinho. E.. and A.R. Cordeiro. 1975. Codominant isoenzymic alleles as
markers of genetic diversity correlated with heterosis in maize (Zea mays). Theor.
Appl. Genet. 46:197-199.

Hotelling. H.. 1954 Multivariate analysis. P 67-80 in O. Kempthorne. T.A. Bancroft.
J.W. Gowan. and J.L. Lush (eds.). Statistics and Mathematics in Biology. Iowa State
College Press. Ames.

Hunter. R.B.. and L.W. Kannenberg. 1971. Isozyme characterization of corn (Zea mays)
inbreds and it's relationship to single cross hybrid performance. Can. J. Gent. Cytol.
13:649-655.

lsleib. T.S.. and J.C. Wynne. 1983. Heterosis in testcrosses of 27 erratic peanut cultivars.
Crop Sci 23:832-841.

Kahler. AL. and C.F. Wehrhahn. 1986. Associations between quantitative traits and
enzyme loci in the F2 population of a maize hybrid. Theor. Appl. Genet. 72:15-26.

Kernpthorne. 0.. 1969. An Introduction to Genetic Statistics. Iowa State University
Press. Ames.

Kiem. P.. RC. Schumaker. and RG. Palmer. 1989. Restriction fragment length
polymorphism diversity in soybean. Theor. Appl. Genet. 77:786-792.

Lamkey. K.R.. AR. Hallauer. and AL. Kahler. 1987. Allelic difi'ercnces at enzyme loci
and hybrid performance in maize. J. Hered. 78:231-234.

Lee. M.. 8.8. Godshalk. ILR. Lamkey. and W.W. Woodman. 1989. Association of
restriction fragment length polymorphisms among maize inbreds with agronomic
performance oftheir crosses. Crp Sci. 29:1067-1071.

Lefort-Buson. M.. Y. Datee. and B. Guillot-Lemoine. 1987. Heterosis and genetic
distance in rapeseed (Brassica napus L): use ofkinship coeflcient. Genome 29:11-18.

Mahalanobis. P.C.. 1936. On the generalized distance in statistics. Proc. Nat. Inst. Sci.
India. 2:49-55.

19

Malecot. G.. 1948. Les Mathematuues de L'heredite. Masson and Cic. Paris

Maluf. W.R. P.E. Ferreira. and J.E.C. Miranda. 1983. Genetic diversity in tomatoes and
it's relationship with heterosis for yield in F1 hybrids. Rev. Brasil Genet. 6:453-460.

Nei. M.. 1965. Variations and covariation of gene frequencies in subdivided
populations. Evol. 19:256-258

Nei. M.. 1972. Genetic distance between populations. Am. Nat. 106:283-292.

Nienhuis. J.. T. Helentjaris. M. Slocum. B. Ruggero. and A. Schaefer. 1987. Restriction
fragment length polymorphism analysis of loci associated with insect resistance in
tomato. Crop Sci. 27:797-803.

Osborn. T.C.. D.C. Alexander. and J.F. Fobes. 1987. Identification of restriction
fragment length polymorphisms linked to genes controlling soluble solids content in
tomato fruit. Theor. Appl. Genet. 73:350-356.

Pearson. K.. 1926. On the coefficient of racial likeness. Biometrika 18:105-1 17.

Prasad. S.R.. and T.P. Singh. 1986. Heterosis in relation to genetic divergence in maize
(Zea mays L.). Euphytica 35:919—924.

Price. S.C.. AL. Kahler. AR Hallauer. P. Charmley. and DA Giegel. 1986. Relationship
between performance and multilocus heterozygosity at enzyme loci in single-cross
hybrids ofmaize. J. Hered. 77:341-344.

St. Martin. S.R. 1982. Efi'ective population size for the soybean improvement program
inmaturitygroups 00 to IV. Crop Sci. 22:151-152.

Shamsuddin. AKM.. 1985. Genetic diversity in relation to heterosis and combining
ability in spring wheat. Theor. Appl. Genet. 70:306-308.

Sokal. RR. 1961. Distance as a measure of taxonomic similarity. Syst. 2001. 10:70-79.

Sokal. RR. 1965. Statistical methods in systematics. Bio. Rev. Camb. Phil. Soc.
40:337-341.

Sokal.RR. and P.H.A. Sneath. 1963. Principles of anerical Taxonomy. Freeman.
San Francisco.

Souza. E.. and ME. Sorrels. 1989. Pedigree analysis of North American oat cultivars
released from 1951-1985. Crop Sci. 29:595-601.

Spiess. EB. 1977. Genes inPopulations. John Wiley and Sons. New York.

Stuber. C.W.. M.D. Edwards. and J.F. Wendel. 1987. Molecular marker-facilitated
investigations of quantitative trait loci in maize. 11. Factors influencing yield and it's
components. Crop Sci. 27:639-648.

Tanksley. S.D.. H. Medina-Film. and CM. Rick. 1982. Use of naturally occurring
enzyme variation to detect and map genes controlling quantitative traits in an
interspecific backer-ass of tomato. Hered. 49:1 1-25.

Vallejos. C.E.. and SD. Tanksley. 1983. Segregation of isozyme markers and cold
tolerance in an interspecific backcross of tomato. Theor. Appl. Genet. 66:241-247.

20
Wright. S. 1917. Coefficients of inbreeding and relationship. Am. Nat. 51:545-559.

Estimation of the Genetic Structure
of an mite Soybean Population

Abstract

The source and frequency of genes. as well as the genetic
distance between lines. was estimated with marker loci data and
coefficient of parentage calculations for a set of 62 elite public and
private soybean cultivars and breeding lines. Through CP and
correlation analysis it was determined that genetic base of the elite
population was derived from a total of 27 ancestral parents with the
parents "Lincoln". "Mandarin Ottawa”. "AK. Harrow". and "Richland"
contributing 30.6%. 17.7%, 10.6%. and 7.9% of the parentage
respectively The marker loci analysis revealed no more than two
alleles per locus within this population and the average probability of
sampling an allele from one line that is identical by state to an allele
sampled from another line was 0.64 while the average CP between any
two elite lines was 0.26. While the data on average suggested a very
narrow genetic base. pairs of elite lines could be identified that
theoretically shared few genes and certain lines were identified as
outliers. There was a poor correlation between the marker loci and
CP estimates of pairwise distances and the overall relationship among
the elite lines. Each method has it's own biases and it would be
valuable to evaluate their accuracy in estimating the actual genetic
distance between individuals in conjunction with unbiased estimates of
the same. Selection of lines for inclusion in a breeding population
based on a narrow range of maturity and yield performance criteria
appeared to result in a more inbred population when compared to a
general set of public cultivars released during the same time frame as
the elite lines that had a broader maturity range and an average CP of
0.19 . Selection of lines for the elite population based on similar
maturity could explain the restricted nature of the elite population
versus the general population. rather than selection of lines for their
similar high yield potential in Michigan.

Introduction

Breeders of autogamous crops tend to use parental lines that already
have an acceptable phenotype for the trait to be improved. The mean of the
progeny (u') of lines selected from a segregating population expressed in
terms of gain from mass selection is equal to the mean of that population
(u) plus the heritable portion of the selection difl'erential.

u'=(u3-u)h2+u

The use of superior parents, termed elite lines, generally assures a high u.
A problem can arise with this approach when the selection of elite parental
lines with similar phenotypes, for example high regional yield and adapted
maturity, results in the selection of similar genotypes. This can reduce the
amount of genetic variability that will be generated in segregating
populations derived by crossing among elite lines and consequently reduce
heritability and gain from selection.

This problem may be more pronounced in crops such as soybeans
(Glycine max L. Merr.) where the adapted gene pool from which elite lines
are derived has a narrow genetic base. Delanney et al. (1983) analyzed the
pedigrees of 110 United States and Canadian cultivars released &0m 1971 to
1980 from maturity groups 00 to IV and found that 50 plant introductions
accounted for all the parentage, 10 of these plant introductions accounted
for 88% of the pool. and two of these, "Mandarin" and "Mandarin Ottawa",
made up 30% of the parentage. The southern gene pool had an even more
restricted genetic base. St. Martin (1982) and Cox et al. (1985a) calculated
an average coefiicient of parentage among northern cultivars released in
the 19808 of .25 and .19 respectively again illustrating the high degree of
relationship among the adapted northern cultivars. It is likely that this
gene pool is even narrower than this data indicates due to probable genetic
relationships among the plant introductions as suggested by their similar
geographic origins (Delanney et al., 1983) and similar isozymes (Cox et al.,
1985) and restriction fragment length polymorphism patterns (Kiem et al.,
1989).

While soybean breeders have recently been using new plant
introductions in their breeding programs to broaden the genetic base. the

2 1

22

incorporation of such germplasm into the elite gene pool often takes
multiple cycles of breeding and a breeder must continue to work within the
existing elite population to develop agronomically superior cultivars. Thus
it would be useful to estimate the genetic diversity within the elite soybean
population and the genetic distance between it's members to maximize the
use of the existing genetic diversity of this population and to hopefully
maximize the breeding progress attained from this population. These
parameters together can be termed the genetic structure of a population.

There are several ways to estimate the genetic distance between
genotypes, each with it's own biases and assumptions. There are extensive
pedigree records for adapted soybeans which can be used to calculate the
coefficient of parentage (CP) (Kemptborne, 1969) between lines; a probability
of genetic identity derived from assumptions of ancestral genealogy,
inbreeding, and gene transmission. Another approach to estimating
genetic distance is to use highly heritable phenotypes, such as isozymes,
restriction fragment length polymorphisms, and certain morphological
traits, as qualitative markers to assay for differences between genotypes at
the DNA level. These polymorphisms occur with a moderate frequency in°
adapted soybeans (Apuya et al., 1988; Doong and Kiang, 1987; Gorman,
1983; Kiem et al. 1989) and the accuracy of the distance estimate will depend
on obtaining a representative sample of potentially polymorphic loci. These
two approaches can also estimate allele and genotype frequencies within a
population. A third approach infers a genetic distance from a statistical
distance that is based on quantitative trait differences (Goodman, 197 2)
would seem less applicable in the narrowly defined elite soybean population
where the range of phenotypes is quite restricted. Furthermore, each
quantitative phenotype assays an unknown number of genetic differences
that may not be proportional to the phenotypic difi'erence and the efi‘ect of
the environment on these traits requires that the phenotype of each
individual be measured in multiple locations and years.

This research was undertaken to (i) see whether selection of elite
parental lines from the northern adapted soybean gene pool based on
regional performance and maturity resulted in a further restriction of the
gene pool and (ii) to determine and compare the genetic structure of a set of
elite lines using the CP and marker loci approaches.

2 3
Materials and Methods

CoefidentofParentegeData

Sixty-two public breeding lines, public cultivars, and private cultivars
(Table 1) were selected for inclusion in this study based on their
performance over multiple years and locations in Michigan State
University yield trials and were collectively termed the elite population.
Forty-seven of the elite lines were in maturity group II, 14 were maturity
group I and one was a group III maturity. Another set of lines termed the
general population (Table 2) was formed from all maturity group 00 to
group IV cultivars that were released after 1979 or that were used as a
check variety in the USDA Northern Uniform Trials conducted from 1980 to
1990. This set was used to represent a broader population of adapted
germplasm developed during the same time frame as the elite lines. All
the elite public cultivars were also in this set.

The CP between two genotypes, defined as the probability that a
randomly sampled allele from one genotype is identical by descent to a
randomly sampled allele from the same locus from the other genotype, was
calculated according to the formula of Kempthorne (1969)

ULY = 1/2 (TLA‘l' 1Y3)

where my is CP between genotypes X and Y, genotypes A and B are the
parents of genotype Y, and m and 11,13 are the CPs between genotype X and
genotype Y's parents. The genotypes ”Elgin" and "Kenwood" were derived
from Iowa State University soybean population AP6 that was formed by an
essentially random mating of 40 cultivars (Fehr and Ortiz, 1975). The CP
between any genotype and one of these two lines was calculated as

'h
m = 1/40 (firgpi)
where m is the CP between genotype X and a genotype derived from AP6
and rm); is the CP between X and the ith parent of the APG population.
Furthermore, two of the 40 parents of APG were derived from a University of
Maryland population formed by a random mating of eight parents and the
CP between these two lines and any other lines was calculated in a fashion

24

Table 1. Summary of the source and name of the 62 elite lines. the average.
minimum. maximum. standard deviation. and correlation of the coefficient of
parentage (CP) and similarity index (SI) values for each line as compared to the
other 61 lines.

 

 

 

 

 

 

 

 

 

 

CP SI
Source Name Avg. Min. Max. Std Avg. Min. Max Std. Corr.
Pub. cultivars BSR 101 .27 .14 .52 .07 .65 .45 .87 .08 .01
838 201 .30 .09 .60 09 .68 .37 .84 .10 .15
BURLISON .21 .11 .59 09 .62 .37 .87 .11 .33**
CENTURY .29 .12 .72 14 .68 .40 .95 .13 .37**
CHAPMAN .30 .10 .47 09 .69 .45 .95 .11 --05
CONRAD .21 .09 .38 06 .66 .40 .87 .09 .43**
CORSOY .25 .01 .94 .16 .65 .50 .97 .09 .54**
ELCIN .24 .11 .64 09 .70 .53 .95 .10 .45**
HACK .31 .04 .71 .12 .58 .42 .74 .08 .18
EARDIN -25 -04 -94 .15 .64 -47 -97 -10 -54**
KENWOOD .28 .08 .64 .11 .70 .48 .95 .11 .09
PELLA .32 .07 .63 .13 .64 .34 .90 .13 .51**
SIBLEY .21 .08 .43 .08 .55 .37 .76 .09 .23
STURDY .25 .12 .60 .09 .67 .53 .87 .08 .27*
ZANE. .31 -10 -63 .08 -69 -47 .95 .09 .33**
Ill. A.E.S. LN82-296 .29 .14 .70 11 .68 .45 .95 .11 .46**
Ind. A.E.S C1664 .24 .13 .56 08 .67 .45 .90 .10 .40**
Iowa A.E.S A80-147002 .32 .09 .63 10 .61 .42 .82 .10 .42**
A82-161035 .29 .12 .50 .07 .61 .50 .90 .08 .14
283e271010 -28 -12 -567 -07 .63 -45 -90 .09 -05
A84-185032 .22 .14 .42 .05 .62 .42 .87 .11 -.25*
ABS-192034 .31 .08 .68 .10 .67 .45 .87 .09 .36**
ABS-291010 .30 .08 .73 .13 .69 .45 .90 .09 .14
ABS-292023 .28 .09 .58 .10 .58 .37 .79 .08 .02
ABS-293033 -28 .13 .45 -08 -58 .37 -90 -11 -08
A86-102004 .28 .10 .63 .10 .66 .47 .92 .11 .20
A86-103017 .24 .12 .60 .07 .65 .42 .84 .10 .14
A86-103027 .31 .06 .71 .13 .61 .42 .76 .08 .29*
A86-202026 .24 .12 .60 .07 .64 .42 .82 .10 .22
A87-198005 32 -09 .256, .09 .64 .40 -87 .09 -34fi*
A87-296011 .31 .08 .56 .10 .66 .47 .87 .07 .06
A87-297015 .19 .11 .42 .07 .61 .42 .76 .08 .03
Mich. A.E.S 884098 .22 .09 .51 .09 .65 .42 .90 .11 .25*
£84150 .24 .11 .54 .08 .67 .50 .84 .09 .32**
E84159 -23 2.07 -59 -08 .60 .37 .79 -09 -36**
£84165 .23 .07 .59 .08 .64 .40 .82 .09 .34‘*
885100 .25 .06 .45 .09 .60 .40 .87 .09 .47**
885166 .29 .15 .61 .08 .61 .40 .79 .10 .49**
885168 .29 .15 .61 .08 .65 .40 .87 .09 .28*
£86339 -26 -12 -62 -08 -67 .50 .87 2.09 -38**
886348 .29 .12 .62 .09 .68 .45 .87 .09 .24
887223 .24 .09 .59 .09 .63 .42 .76 .08 .22
888080 .28 .14 .63 .09 .68 .47 .84 .09 .21
Minn. A.E.S. M82—946 .23 .13 .38 .05 .56 .34 .76 .09 .25*

 

 

 

 

25
Table 1. Continued

 

 

 

 

 

 

CP SI
Source Name Avg. Min. Max. Std. Avg. Min. Max. Std. Corr.
M84-916 .25 .10 .37 .06 .62 .40 .82 .10 .39**
Ohio A.E.S. HW8008 .26 .13 .72 .12 .65 .45 1.00 .12 .37**
Callahan Seeds 1250 .23 .14 .42 .06 .61 .37 .90 .12 .16
7260 .26 .13 .40 .06 .64 .37 .87 .09 .03
1299 .15 .09 .35, 2.06 .61 .42 .82 .09 2.17
8252 .28 .13 .61 .08 .60 .37 .79 .10 .26*
Pioneer 9271 .24 .09 .50 .07 .64 .42 .92 .10 .26*
9292 .24 .09 .50 .07 .66 .40 .92 .10 .24
Asgrow Seeds A1937 .30 .06 .73 .14 .67 .47 .92 .10 .28*
A2234, -30 -.01 -61 .13 .65 -45 1.00 -12 -18
A2943 .19 .10 .35 .04 .60 .37 .74 .08 .04
Agripro AP 1989 .26 .10 .57 .09 .59 .45 .87 .09 .37**
Dairyland Seed USE-262 .24 .06 .60 .09 .67 .40 .84 .11 .29*
Jacques Seed J-231 .26 .09 .66 .10 .68 .53 .90 .08 .24
Northrup King 31884 .30 -08 -59 -10 .61 -42 .82 .08 251**
819-90 .18 .04 .51 .09 .60 .37 .79 .11 .31*
823-12 .13 .06 .31 .05 .59 .37 .79 .10 -.02
32596 .22 .09 .31 .05 .68 .47 .90 .11 .28*
Over all comparisons .26 .01 .94 .08 .64 .34 1.00 .10 .27*'

 

 

 

 

Corr. = correlation between the coefficient of parentage and similarity index for
that elite line.

'. “ denote significant correlations at the 0.05 and 0.01 levels of probability
respectively.with 59 df.

26

 

 

 

Table 2. Summary soybean cultivars included in the general
population.
Maturity Group
IV III II I 0 00
Avery Bass Amcor BSR 101 Aries Bicentennial
Douglas BSR 302 Amsoy Hardin Chico Maple Amber
Eygptian Cartter Beeson Hodgson Clay Maple Arrow
Flyer Chamberlain Bell Kasota Dassel Maple Isle
Franklin Cumberland BSR 201 Kato Dawson Maple Ridge
Hamilton Edison Burlison Lakota Evans McCall
Lawerence Fayette Century M82-106 Glenwood Portage
LN83-2356 Fremont Chapman M83-899 Libra Proto
Morgan Harper CN210 Sibley Maple Donovan Scorpio
Pennyrile Hayes CN290 Sturdy Minatto
Pershing Jack Conrad Weber Ozzie
Pyramid Logan Corsoy Piceses
Regal Mead Elgin Simpson
Sparks Pella 886339
Spencer Resnik Hack
Stafford Sherman Kenwood
Union Williams Marcus
Winchester Miami
Zane Newton
Platte
Preston
Vinton
Wells
17a 19 23 11 13 9

 

a number of genotypes listed in the column.

27

similar to that of the AP6 derived lines. The following assumptions were
made in the CP calculations: (i) all plant introductions are completely
unrelated, including lines selected from other plant introductions, (ii) all
plant introductions were completely inbred, and (iii) that each parent of a
biparental cross contributed equally to all progeny derived from the cross.
A genotype derived by five or more backcrosses was considered to be
genetically equivalent to the recurrent parent. The pedigree information
used in the CP calculations was obtained from public sources and in a few
instances, the private companies that developed the genotype.

The CP between an elite line and an ancestral parent is the
probability that the elite line carries an allele from that ancestral parent at
any particular locus but it can also be thought of as the percentage of the
elite line's genes that derive from that ancestral parent. The CP between
an elite line and an ancestral parent, hereafter referred to as the ancestral
parent contribution, were calculated between each elite line and each
ancestral parent found in the pedigrees of the 62 elite lines, producing a
multivariate data set describing each elite line with p variables where p is
the number of ancestral parents. This data set was used to estimate the
genetic constitution of the elite population and to examine the
interrelationship of the elite lines through principal component analysis of
the variance-covariance matrix and cluster analyses. The data was also
analyzed for correlations between the ancestral parent contributions. All
CPs were calculated using computer programs written by the author in the
SAS Interactive Matrix Language. All other analyses were performed
using SAS software.

MarkerLodDaia

Ten seeds of each of the 62 elite lines were tested for their genotypes at
the following thirteen isozyme loci: diaphorase (Dia1 locus; EC 1.6.4.3),
endopeptidase (Enp locus), isocitrate dehydrogenase (Idh1, Idhz loci; EC
1.1.1.42), mannose phosphate isomerase (Mpi locus; EC5.10.11),
phosphoglucosemutase (Pgm1 locus; EC 5.3.1.9), and superoxide dimutase
(Sod locus; EC 1.15.1.1) using the D buffer system of (Cardy and Beversdorf,
1984) and acid phosphatase (Acp locus; Ec 3.1.3.2), aconitase (Acoz, Ac04
loci; EC 4.2.1.3), fluorescent esterase (Fle locus), 6-phosphogluconate

28

dehydrogenase (Pgd1 locus; EC 1.1.1.44), and phosphoglucose isomerase
(Pgi locus; EC) using a Tris-Citrate system (.2 M Tris, .062M citric acid, pH
6.7 electrode bufl'er/ .0075 M Tris, .002 M citric acid, pH 6.7 gel bufi'er). All
gels were 11.5 % (w/v) starch. An additional ten seeds were analyzed if
heterogeneity was found within a line.

To prepare electrophoretic samples, seeds were imbibed in distilled
water for 24 hours after which half of one cotyledon was removed and
ground in 0.25 ml of grinding buffer (16.6% (w/v) sucrose, 8.3% ascorbic
acid, pH 7.4). D system gels were run for 6 hours at 275 volts/50 mAmps
while the Tie-citrate gels were run for six hours at 200 volts/50 mAmps.
The stains and procedures of (Cardy and Beversdorf, 1984) were used for all
isozymes except florescent esterase which was resolved with a stain using
40 ml of 0.1 M sodium acetate-HCl bufl'er (pH 5.2) and 15 mg of 4-
methylumbelliferyl acetate, dissolved in 10 ml of acetone. Fle Bands were
then visualized under longwave UV light.

The genotype was determined for each elite line for the W, T, L, I,
and R loci that control the color of the flower, pubescence, pod, hilum and
hilum respectively. The level of peroxidase activity was assayed using the
method of Buttery and Buzzel (1968) under the assumption of single gene
inheritance of this polymorphism (Buttery and Buzzel, 1968).

No more than two alleles were found for any locus among the 62 elite
lines such that the frequency of one arbitrarily chosen allele (pi) could be
used to represent the genotype of a line for the jth locus in the marker loci
data set . Loci that exhibited heterogeneity within a line were assigned a p;
of 0.5 under the assumption that the heterogeneity was a result of
heterozygosity at that locus in the single plant from which the line was
bulked and that assaying a larger sample would have produced an actual p;
of 0.5 (Appendix A). A similarity index (SI) equivalent to the probability
that an allele randomly sampled from one genotype is identical by state to
an allele randomly sampled from the same locus from the other genotype
was calculated between all elite lines. The SI with this data was

r. n
SI = (1/11) 2 2 Pijx 9in

"l J“

where pij, and pijy are the frequencies of the ith allele at the 5th of n loci in
genotype X and Y respectively. The SI is equivalent to the I, measure

29

reported by Sokal and Sneath (1963) and is equivalent to the numerator of
Nei's index before normalization (N ei, 1972). The allele frequency data was
also used in principal component, principal coordinate, and cluster
analyses. The SI and principal coordinate analyses were performed using
a programs written by the author in SAS Interactive Matrix Language

while principal component and cluster analyses were performed with
standard SAS software.

Results
CoeficientofPamntage

The average CP between the elite lines was .26 (Table 1) (Appendix B)
and pairwise CPs ranged fi-om a low of 0.01 between "Corsoy” and A2234 to
a high of 0.94 between "Hardin" and it's recurrent parent in three
backcrosses, Corsoy. Some lines were more related to the elite population
as a whole than others as shown by the average CP for individual elite lines
that ranged from 0.13 for Northrup King s23-12 to 0.32 for A87-198005 and
”Pella" (Table 1). Every elite line could be paired with at least one other elite
line to produce a CP of less than 0.16 . Removing the five elite lines that
appeared to be least related to the population fiom the data set increased the
average CP between the remaining elite lines to 0.27 (Table 3).

The average CP between the general set of group 00-IV maturity lines
was 0.19 (Table 3). Subdividing this population by maturity groups showed
that the the average CP between lines within maturity groups was about
equal to that between the elite lines and that the bulk of the diversity among
these lines was distributed between lines with different maturities (Table 3).

The use of multivariate statistic techniques facilitated the
investigation of the overall relationship between the elite lines. While
cluster analysis could have been done on the 62 x 62 matrix of CPs between
the elite lines, a more informative analysis was performed by using the
multivariate ancestral parent contribution profile of the elite line as this
allowed an analysis of the genetic base of the elite population and the source
of the variation between the elite lines.

This analysis showed that 34 plant introductions contributed genes to
the elite population (excluding the donor parents of lines derived fiom five

30

Table 3. The average coefficient of parentage (CP) between lines of
the general population and it's maturity subdivisions and between the
elite lines with or without a subset of the elite lines that appeared to
be outliers.

 

 

Population CP na

General OO-IV .19 92

O-III .22 66

I-III .25 53

I-II .24 34

00 .15 9

0 .24 13

I .27 11

II .24 23

III .35 19

IV .25 17
Between general lines of same maturity .27
Between general lines of different maturity .17

Elite lines .26 62

Elite lines - outliersb .27 56

 

a number of lines. in the population or subdivision.

b outlier were the elite lines $19-90, 823-12, 7299, A87-297015,
Corsoy, and Hardin.

31

or more backcrosses) with 15 of these contributing at least 1% of the
parentage (Figure 1) while the other 19 contributed a total of only 3.8% of the
parentage. The first twelve plant introductions accounted for 91.4% of the
parentage. There were many significant positive correlations (Table 4)
between the plant introduction contributions to the elite population that
could be traced to the correlated plant introductions being used primarily in
unison as parents in only one cross. This means that the genes fi'om the
correlated ancestors shared a common pedigree as a route of introduction
to the elite gene pool. Coupled with progeny inbreeding, this represents a
further restriction on the elite gene pool as an inbred genotype can pass on
to it's progeny an allele from only one of it's parents. For example, the
contributions from "Mandarin" and "Manchu" were perfectly correlated
and each was used only once in forming the elite population, as the parents
of the cultivar "Lincoln". As an inbred line, Lincoln can pass on to
subsequent progeny and elite lines only one type of allele, derived from
either Mandarin or Manchu. Thus the contributions from Mandarin and
Manchu to the elite population can be summed and appropiately termed the
contribution from Lincoln in recognition of Lincoln as the sole source of
these genes.

All ancestral parent contributions with correlations of unity were
pooled as were the contributions fi'om "Illini" and "Dunfield" (r = 0.99) and
termed the contributions from the primary derived progeny (Table 4).
While not pooled, the contributions from "No. 171" and "A.K. Harrow" also
followed a restricted entry route into the elite population. Both were parents
of "Capital” which is a parent of Corsoy, while A.K. Harrow (along with
"Mandarin Ottawa") was also a parent of ”Harosoy" which in turn was the
other parent of Corsoy. The major source of entry of Harosoy genes to the
elite population is through the extensive use of Corsoy as a parent
indicating a common entry route for the genes of No. 171 and A.K. Harrow.

Figure 2 shows the genetic constitution of the elite population after
the pooling of correlated ancestral contributions which resulted in the
parentage of the elite population being effectively derived from only 27
ancestral parents or their derived progeny with Lincoln accounting for
30.3% of the parentage and the first nine genotypes accounting for 91.4% of
the elite population's parentage.

32

 

I
Mandarin 0t:

 

muu. _

-:::l=%'—

liable-d

Io. 111
.lukdea

Ancestral parent
3
2

 

lllllLlllllllllllL

Cl General
1 m I Elite
10 15 20

 

96 of Parentage

Figure 1. The percentage of the elite and general populations's
parentage that derived from the ancestral parents that contributed at
least 1.0% of the parentage.

Table 4.

33
Summary of significant correlations between the

contributions from ancestral parents (APs) to the elite population (re)
and the broader population (Too-Iv). their contributions to the elite
population (APC) along with the primary progeny line derived fi'om the
correlated ancestral parents.

 

 

Progeny

APl APZ re too-Iv APCla APC2 Total Line
Mandarin with Manchu 1.00 0.00 15.30 15.30 30.60 Lincoln
Tokyo with PI 546.10 1.00 0.00 2.95 2.95 5.90 Ogden
Patoka with PI 840.41 1.00 0.00 .40 .40 .80 Perry
FC31745 with PI 171.442 1.00 0.00 .13 .06 .19 Tracy
FC33243 with C143 1.00 0.00 .09 .04 .13
FC33243 with Mansoy 1.00 0.00 .09 .04 .13
Mansoy with C143 1.00 0.00 .04 .04 .08
Flambeau with Pagoda 1.00 0.00 .04 .03 .07
Illini with Dunfield .99 0.00 3.02 3.12 6.14 Adams
A.K. Harrow with No. 171 .94 0.00 10.55 4.81 15.36 Capital
Roanoke with PI 257.345 .93 0.00 1.26 1.61 2.87 NKs1346
$100 with Patoka .89 0.00 .72 .40
$100 with PI 84041 .89 0.00 .72 .40
Richland with Mukden .86 0.00 7.88 4.74
Haberlandt with PC 31745 .84 0.00 .10 .13
Haberlandt with PI 171.442 .84 0.00 .10 .06
Mandarin Ottawa with A.K. Harrow .74 0.00 17.67 10.55 Harosoy
Ralsoy with Patoka .72 0.00 .05 .40
Ralsoy with P1 840.41 .72 0.00 .05 .40
Mandarin Ottawa with Manchu -.71 0.00 17.66 15.30
Mandarin Ottawa with Mandarin -.71 0.00 17.66 15.30

 

 

 

 

a APC 1 and APC2 are the ancestral parent contribution to the elite
population from ancestral parents 1 and 2 respectively.

34

 

l I l I F I

Lincoln -

Mandarin Ott -

A! Marrow ..

4') Richland ..
t:

0 Adana ..
H

a Ogden ..
d

H Mukdon -
U

3 cu -
0

g P! 180.50]. -
'1:

PI 257.345 -

Roanoke -

8100 ..

l l l i L I

 

 

’6 of Parentage

Figure 2. The percentage of the elite population's parentage that
derived from the ancestral parents that contributed at least 1.0% of
the parentage.

35

A principal component analysis was performed on the variance-
covariance matrix of the ancestral contribution data set to summarize the
source of the variability among the elite lines. The first three principal
components accounted for 86.4% of the variability among the elite lines and
the position of the elite lines relative to these coordinates is shown in Figure
3. The first principal component accounted for 56.9% of the variability and
the correlation of the original ancestral parent contributions with the
scores for this principal component (Table 5) indicated that it primarily
seperated the elite lines based on what proportion of their genome
originated from Lincoln or from A.K. Harrow, Mandarin Ottawa, and No.
171 which are the parents of Harosoy and Corsoy. This showed that a
major source variation among the elite lines can be attributed to what
proportion of their parentage derived from these four ancestral parents that
combined accounted for 58.6% of the elite parentage. The second principal
component separated lines primarily based on their genomic content from
the ancestral parents "Roanoke"'and PI 257.345, or from No. 171 and A.K.
Harrow while principal component three separated the elite lines primarily
based on the proportion of their genome derived from "Ogden", "Richland",
and "Mukden" or from Roanoke and PI 257.345. PI 257.345 was only found
in the pedigree of Northrup King 81346 (along with Roanoke) which in turn
is one of the parents of only four elite lines.

Cluster analysis can incorporate all dimensions of a multivariate
data set in it's resulting dendogram while principal component analysis
can only graphically present three dimensions at best. A cluster analysis of
the ancestral parent contribution data set was performed using various
linking methods using only those variables found to significantly contribute
to the variability of the data set through the principal component analysis.
The results showed that while cluster membership was fairly constant
across linking methods, this being especially true for tight groupings, there
was considerable variation regarding the position of the tight groups and
outlying genotypes relative to one another. Thus the association between
many of the elite lines became dependent on the biases of the linking
method and it appeared to be a very subjective method of determining the
genetic distance between any two elite lines. The cluster analysis seemed to
have more utility in graphically presenting an overall relationship between
all the elite lines and only the clustering produced using Ward's linking

36

.Av whoofim oomv coanouwuo mcwuoumoao m.oum3 one memo meow one means mocfla oufiao
onu mo ocfluoumoao Havanououmflz sown oo>finoo mnoumoao o>flu man on oGOQmouuoo moamnm .mmcfla
ooze mm on”. no oumo counuonfluucou ucoumn Hmuumoucm on» no mwmcfimcm unocoasoo Hmawucflun
on”. 80.3 340m one .mdum .28.: 3550960 16325.3 manna you: on» «0 Doom .n own—3E

on d? m...

 

«06
EN .

o ~..~..

. v0 .0:

and

vnfi

mg...
m you-moan 0
v nouns-Ho "A
m Hon-moan unn—
m woumsfiu u D
H nouns-Ho no

37

Table 5. Correlations of the contributions from ancestral parents that
were a source of at least 1% of the genes in the elite population with
the principal component scores fi'om the analysis of this data set along
with the the percentage of the variation in this data set that is
accounted for by the first three principal components.

 

Correlation with
Principal Component

 

 

Ancestral Parent APCa 1 2 3
Lincoln 30.6% -.97** -.22 .03
Mandarin Ottawa 17.7% .84** -.34** .36**
A.K. Harrow 10.6% .71** -.65** .21
Richland 7.9% -.14 .52** .52**
Adams 6.1% -.36** .22 .00
Ogden 5.9% .14 .36** .69**
No. 171 4.8% .52** .70** .46**
Mukden 4.7% .36** .58** .53**
CNS 3.4% -.56** .09 .21
PI 180.501 1.6% -.01 .14 .13
PI 257.345 1.6% .17 .75** -.54**
Roanoke 1.3% .16 .78** -.56**
% of total variation accounted

for by the principal component 56.87% .91% 10.33%

 

a Ancestral parent contribution to the elite population.

” Denotes a significant correlation at the 0.01 probability level and 60

(if.

38

method which minimizes within cluster variation while maximizing
intercluster variation is presented (Figure 4) as this method seemed to
produce the most logical association between the elite lines based on obvious
pedigree relationships. For example, Corsoy figured prominantly in the
pedigrees of E87223, ”Sturdy”, "Sibley", J-23l, and AP 1989 and Ward's
linking algorthm placed them in the same cluster while other methods
positioned Corsoy as an outlier from all elite lines. The validity of this
structure is also suggested by Figure 3 where members of the same cluster
were grouped together in the graph of the first three principal components
and by an analysis of the ancestral parent contribution profile of each
cluster (Table 6) that showed that the two clusters that were most divergent
from the the rest of the lines were at the extremes in terms of the amount of
their genes that derive from the ancestral parents that were highly
correlated with the first principal component.

Marka'LociData

The average SI value among all pairwise comparisons of the elite
lines was 0.64 (Table 1)‘ (Appendix C) indicating that an F1 obtained by
mating two randomly selected elite lines would be homozygous at 64% of it's
loci. The SI values ranged from a low of 0.34 between Pella and M82-946 to a
high of 1.00 (identical, homogeneous genotypes) between HW8008 and A2234
(Table 1). The average SI value for individual elite lines compared to all
others ranged fi'om a low of 0.55 for Sibley to a high of 0.70 for Kenwood and
Elgin (Table 1). Some of Sibley's uniqueness could be attributed to the fact
that it is the only elite line that carries the Sod allele while the similarity of
Elgin and it's progeny, Kenwood, to many elite lines may be due to their
derivation from the broad based AP6 soybean population. Every elite line
could be paired with at least one other to produce a SI value of 0.53 or less.

No more than two alleles were found among the elite lines for any
locus. The frequency of the allele choosen to represent each locus in the
marker loci data set is presented in Table 7 along with the number of elite
lines that were heterogeneous at the associated locus. Forty-eight percent
of the elite lines were heterogeneous at at least one locus. Fifty-five percent
of the breeding lines were heteogeneous at an average of 1.9 loci per line
while 42% of the released cultivars were heterogeneous at an average of 1.1

 

 

l
’1”!
cluster 1 *ag
ma:
non ‘
an
o-
C“!
m

cluster 2

 

 

 

 

 

 

 

cluster 3

 

 

 

 

 

 

 

 

Figure 4. Results of clustering the 62 elite lines using the ancestral
parent contribution data set after pooling the highly correlated
contributions and using Ward's minimum variance clustering criterion.

‘40

Table 6. Ancestral parent contributions to the five clusters of elite
lines obtained by Ward '3 clustering method.

 

 

Cluster

Ancestral Parent 1 2 3 4 5

Lincoln 11.179 24.81 35.66 14.21 50.36
Mandarin Ottawa 28.70 21.65 15.35 9.96 10.61
AK-Harrow 24.29 8.53 10.28 4.88 6.32
Richland 4.74 10.22 6.42 10.21 8.63
Adams 1.28 5.88 8.39 5.47 6.51
Ogden 4.49 9.97 4.37 4.69 3.58
Mukden 3.78 8.39 2.44 8.98 2.67
No. 171 14.75 1.32 5.58 1.56 2.78
PI 257.345 0.00 0.00 0.00 25.00 0.00
CNS 0.96 2.94 4.32 0.98 5.08
Roanoke 0.49 0.70 0.50 12.50 0.00
PI 180.501 3.16 0.99 0.91 1.56 3.47

 

a Ancestral parent contributions are expressed as percentage of the
cluster's genome.

41

Table 7. Summary of allele frequencies and heterogeneity in the elite
population for nineteen marker loci and the correlation of the
frequencies with the first three principal components.

 

 

Allele Principal Component
Trait Allele frequency na 1 2 3
Flower color W1 .73 4 -.17 -.SS** .45**
Pubescence color Cl .45 0 .67** .24 .16
Pod color br .67 3 .61** -.12 -.11
Hylum color 1 .74 0 -.50** .18 -.27*
Hylum color r .52 0 .42** .19 -.37**
Idhl .75 5 -.21 .17 .02
Idhz .41 2 ‘u47** .05 -.01
Ac04 .60 1 .09 —.17 -.11
Acoz .95 0 -.21 -.06 -.06
Mpil .35 4 .25* .58** .12
Pgml .40 5 .18 -.02 -.77**
Dial .65 3 -.34** -.SS** .33**
Enp 1.00 0 .00 .00 .00
Per .35 3 .64** -.10 .33**
Sod .02 0 .09 .24 -.15
Fle .36 12 .46** -.16 .33**
Pgi .53 4 -.19 .76** .33**
Aop 1.00 0 .00 .00 .00
Pgdl .84 2 .17 -.17 .06

 

‘. “ denote significant correlations at the 0.05 and 0.01 probability
levels respectively with 60 df.

a number of elite lines that were heterogeneous at the locus.

42

loci per line. The high frequency of heterogeneous loci in these supposedly
inbred lines may be attributed to the fact that half are breeding lines that
have not yet gone through a final purification and that even then,
biochemical homogeneity is rarely ascertained nor selected for. Enp and
Acp were monomorphic in the elite population while Sad and Acoz, were
represented by an alternate allele in only one and two elite lines
respectively. '

The variance-covariance matrix of the allele frequency data was used
in a principal component analysis that resulted in the first three principal
components accounting for only 39.0% of the variation, indicating that
graphing the observations in these three dimensions would not produce a
reliable picture of their actual relationship. The correlations of the marker
loci to the principal componets (Table 7) did not appear to produce a clear
interpretation of the new axses. To try and achieve a better data summary,
a principal coordinate analysis was performed on a 62 x 62 distance matrix
between the elite lines with 1 - SI being the distance measure. This
multivariate technique attempts to produce a low-dimensional plot in
Euclidian space where the proximity of the points approximates their
original distance. The first three principal coordinates accounted for 59.1%
of the variation and the postion of the elite lines relative to these coordinates
are shown in Figure 5. The elite lines did not appear to form mutiple
groups in this space, rather they were simply dispersed around the
centroid with a few outlying elite lines. Cluster analysis of this data also
failed to produce distinct groupings within the elite population that was
independent of what linking method was used. The dendogram that
resulted from the average linking of elite lines is shown in Figure 6. It's
validity seemed to be supported by a similar association of points in the
principal coordinate graph (Figure 5).

The correlation of the CP and SI values between pairs of elite lines
was 0.27 (Table 1) and while this is highly significant with 1879 df, it
indicated that one measure was not very predictive of the other suggesting
that each measured genetic distance independently. For individual elite
lines compared to all others, the correlations ranged from -0.25 for A84-
185032 to 0.54 for Corsoy and Hardin (Table l). The reason for the variation
is not apparent. More evidence for the independence of these distance
measures came fi'om representing the points in the principal component

43

.Am eunofim memo nOflneuwno unencumnao emene>e on» one new eueo oEem on» onfimn
menaa euHHe enu mo onwueumnau Heoflnouenewn Eowu oe>wneo mueumnao m>flm on» on ononmenuoo
menenm .men: ouflae mm on... no eueo “80H noxuee can no mflmhaene euenwouooo Heaeunwnn
on... noun Amoom one .Noum .Hoom. meuenaonooo Heawonflum counuumufl can no “.on .m cannon

3 .92. on

\\
+ Ill \\\\\ "cum
I \
New .0 III \\\\
II II \\
\
It 0 to \\
A II \\I o
’I’ ‘\\‘ ’ N o
on .o .. ... .
\ III
II
I

‘
\\\\
I \\\
\v/ .

on .0!
III ”N can.

\\ \X” w III m“ 00
\ ‘ I N“ 0°
W. x. no"...
I.. ,x
a. ,,
I

neumnao
neumnau
neumnau
neumoau
neumnao

 

Hmmvm
II II
nooec

44

 

Cl“.
M ’37
nee-unco-

”In! r—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LUZ-29‘
“6.231226 1"-

tum
“9:361? ] '__

3"“

2.2... 7—}-—_. _
LIP
j;

 

 

 

 

 

 

 

 

cluster 1 “an
J-le
as”
””1010
AIS-192339
I.“
mun
m-mcos
can.
“7'2“3'1

. 109-90
Med

m“:3uz 7

m.“ _

M.. W.

(.7213
cluster 2 an . _J

m2

(.913, —

”“7302 l—

oil-fl

__,i§i§ .——.—7_

3.5!
cluster 3 Ala-mm

ans-2mm I t——
___' an: '

tune. —
m __I 1—_
AP 19”
(IMO
1330 —r
us- 293:3! l
Ade-I 03312 -—
7240 —]
72”
’ 0232
A294!

cluster 5 ass-29m: j—l—

roe-9“

 

 

 

 

 

 

 

 

 

 

 

 

 

h-i

 

 

 

 

—r

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

cluster 4

 

 

 

 

 

 

 

 

 

 

 

 

e 6. Results of clustering the 62 elite lines using the marker loci
data set and the average clustering criterion.

45

and coordinate graphs of the ancestral parent contribution and marker loci
data respectively according to the clustering pattern produced by the other
data set. Figure 7 shows the members of a marker loci cluster dispersed
across the space defined by the first three principal components of the
ancestral parent contribution analysis while Figure 8 shows the memebers
of the ancestral parent contribution clusters dispersed across the space
defined by the first three principal coordinates of the marker loci analysis
indicating that the groupings achieved by one data set differed fi'om those
produced by the other data set.

Discussion

The elite lines can be thought of as a selected portion of the soybean
germplasm developed &0m 1980 to 1990 and the CP data suggested that the
elite population was more inbred (CP = 0.26) (Table 1.) than a general group
of cultivars ranging in maturity from group 00 to group IV (CP = 0.19)
(Table 3) that were developed during the same time frame. Cox et al.
(1985a) reported an average CP of 0.19 between maturity group 00-IV
soybeans released fi'om 1971 to 1980 while St. Martin (1982) calculated an
average CP of 0.25 between lines of similar maturity released from 1976-
1980 under the assumption that the ancestral parents Mandarin and
Mandarin Ottawa were indentical. The increased inbreeding of the elite
population appeared to result fi-om the narrow maturity range represented
in the elite population, and not from selection for high yield in Michigan
per se as the average CP between the elite lines was equivalent to that
between the unselected maturity group I-III or group I-II lines (Table 3).
This analysis indicated that the genetic diversity among adapted northern
soybeans was primarily distributed between, rather than within maturity
groups and that many unselected genotypes could be paired to a completely
unrelated line outside of their own maturity group while this occurred only
once when comparisons were made within maturity groups.

The elite soybean population consisted of a set of highly related lines
as shown by an average CP among the elite lines of 0.26 (Table 1), or slightly
higher than that expected between half-sibs derived fi'om unrelated inbred

parents. This means that under the assumptions of the CP that the
average F1 obtained by mating two randomly selected elite lines would be

.3
eunofim com. noaueuwno unencumnau eoeno>e one one new eueo “00H noxuee onu onwmn menfla eufiHe
can no unencumnao Heownonenefin scum oe>wweo mueumnau o>wm enu ou ononmmnnoo memenm .menfia
ooze mm on» no eueo nowunnflnunou unenen Henumeone one no mfimaaene unenoneoo Heafionfinn
on» 50.3 .mnum one .mcom .353 munenonnoo Henwonflnn mews» umnwu one no uon .b manna-m

8 em. ~..

 

% 3 m
hd
L Nvfil
A
vodl
mud
vmfi
«(on
m neumnao "AV
v neumgo "A
m neumnao no
N neumnao "Av
H neumnflo nu

47

. .3
madman eemv nownouanu unencumnao m.owe3 one new eueo nodunnwnunoo unenen Heuumeone on» anew:
men: mow-no on» no unencumoao decenonenewn noun oo>flnoo mneumnau mafia one on ononmennoo
menenm .men: maze am one no eueo flood Movies on... no mwmhaene euenflouoou Henfiunflnn
on» Eonu Amoom one .moum .Hoomv meuenflonooo Henflonfinn eennu umnwu can no uon .m manor"

3 6...... on
mm .0...
I’ll W 0°-
«8.. u xx \\ 89.
I \\\
2 .o ./ \..

I II \ \
II A II \\\ w

e
on 0 Hr 1: I
I
I \
II \
\dl
\\ I
\
\\\
\\ I
\
\
\\

 

I: 0.0
\\ I” 00.0...
. . .. . \. n~.o..
\ .X. m. .o
H . . wad
. . / m8...
m umumnau u 0 %
v neumnau "A P
m neumnao ".0
m neumoao u n—
H neumnao u 0 .

48

homozygous at 26% of it's loci, leaving the remaining 74% to generate the
genetic variability in the succeeding generations of selfed progeny. The
probable genetic relationships among the ancestral parents (Delanney et
al., 1983; Cox et al., 1985a; Kiem et al., 1989) violate the CP assumptions and
would result in a actual genetic similarity between elite lines being greater
than this estimate, perhaps approaching the 0.64 value estimated by the
marker loci based SI which makes no such assumptions (Table 1).

The restricted nature of this genetic base was also illustrated by the
analysis of the ancestral parent contributions that showed that almost 67%
of the parentage derives from only four ancestral lines with Lincoln
contributing 30.6% of the parentage (Figure 2). The significant correlations
between unpooled ancestral parent contributions (Table 4) indicated that the
genetic base is effectively narrower than suggested by the profile shown in
Figure 2 . This profile was remarkably similar to those reported by
Delanney et al. (1983) for maturity group 00 to IV varieties released from
1971 to 1981 and flour 1961-1970 indicating that any efforts in the last ten
years to broaden the genetic base of northern adapted soybeans have yet to
appear in the parentage of this elite population. The CP data only reflects
the frequency that breeders have used the ancestral lines or their derived
progeny as parents in crosses and does not reflect the effects of the breeder's .
selection among the progeny of a cross on the genetic makeup of the elite
population. All CP calculations assumed no selection or sampling and St.
Martin (1982) indicated that when averaged over all genes that the
deviations from the expected value would be small. Yet there certainly
must be some biases toward the genes of one parent over the other,
especially for genes with major effects on a selected trait and these biases
coupled over successive cycles of selection could significantly alter the
genetic constitution of the elite population from that predicted by GP
analysis. Indeed if the ancestral parent profiles have remained relatively
unchanged over the past thirty years then the yield of adapted soybeans
would have remained unchanged without effective progeny selection.

While the average values fiom all analyses of CP and SI data showed
a narrow genetic base, each analysis was also able to detect considerable
variation between the elite lines. Each elite line could be paired with
another elite line to produce a CP and SI value that were considerably lower
than the means (Table 1) and the low average CP and SI of some elite lines

49

compared to all others suggested that certain lines were quite diverse from
the main body of the population. The graphs of the principal component
(Figure 3) and coordinate analyses (Figure 5) performed on the ancestral
parent contribution and marker loci data sets respectively also illustrated
this diversity and showed the existance of outlying genotypes.

The principal component analysis of the ancestral parent
, contributions to the elite popualtion suggested that much of the diversity
among the lines is due to what proportion of their parentage derived fiom
the ancestral parents that were the major sources of genes, particularly
Lincoln and Mandarin Ottawa, A.K. Harrow, and No.171 which are the
parents of Harosoy and it's progeny Corsoy (Table 5). This is illustrated by
the differences in the ancestral parent contribution profiles of the
groupings obtained from Ward's clustering of this data, where the
percentage of parentage derived fi'om Lincoln or from Mandarin Ottawa,
A.K. Harrow and No. 171 are 11.2% and 67.7% respectively for cluster one
and 50.4% and 19.7% respectively for cluster five (Table 6). The isozyme
analysis of the genetic diversity of the elite population found the minimum
amount of allelic diversity indicating that the genetic diversity within this
population had an interlocus versus an intralocus source (Table 7). These
results indicated that apparently diverse genotypes can perform well in
Michigan and that it may be possible to generate variable progeny
populations with high mean yields from which better progeny could be
selected by crossing selected elite lines.

Both the CP and marker loci indicated that gentic diverisity existed in
the elite population yet both appeared to measure the diversity between
individuals independently as shown by the low correlation between the CP
and SI measures (r = 0.25) (Table 1.) and by Figures 7 and 8 where the
grouping of the lines by one data set fails to correspond to the grouping
obtained with the other. Cox et al. (1985a) reported a correlation of 0.63
between the CP and a similarity index based on 20 marker loci fi'om a set of
soybean cultivars released fi'om 1971 to 1981 though the correlation was only
0.48 among lines that were released fi'om 1961 to 1970 and only 0.24 among
lines released before 1961. A low though significant correlation has also
been found between such estimates of genetic similarity in cultivated
Triticum aestivum L. (Cox et al., 1985b). The independence of these
estimates undoubtably arises from the different assumptions and type of

50

data used in the calculations. The CP is a probability based solely on
assumptions and is only as valid as those assumptions, while the SI is a
probability based on actual genetic differences and will be as accurate as the
data and it's assumptions allows. An illustration of the differences
between these measures is provided by the fact that twenty-seven ancestral
parents contributed genes to the elite population (after the pooling of
contributions suggested by the correlation analysis) and the CP assumed
that each contributed an unique allele to the population at every locus while
the marker loci analysis discovered no more than two alleles at any locus
within the elite population (Table 7). Other researchers using isozymes
(Chiang, 1985; Doong and Kiang, 1987; Gorman, 1983; Doong and Kiang,
1987; Chiang, 1985) and RFLPs (Apuya et al., 1988; Kiem et al., 1989) have
also failed to find any allelic diversity approaching the level assumed by the
CP even across Glycine species. The SI on the other hand assumed that
alleles in different individuals that produced the same phenotype are
identical though there can be cyptic variation between them (Ramshaw et
al., 1979) that would cause the SI to overestimate the genetic similarity
between lines. The CP assumed that all progeny fi'om a cross are
genetically equivalent while the SI can reflect the effect of selection and
random sampling on the genotype of the derived lines though it is not
known whether the allele fi'equencies for the elite population reported in
Table 6 reflect these effects or not. The accuracy of the SI estimate of
genetic distance is also dependent on sampling considerations such as the
number of genes in the genome, the number and distribution of the assayed
loci, the size of linkage blocks, and the extent of the linkage disequilibrium
between ML and the loci within these blocks. The SI accuracy could be
improved by assaying more loci and it's relation to genetic diversity for a
particular trait could be further improved by assaying marker loci that are
known to be linked to chromosome segments that afi'ect that trait (Lee et al.,
1989). '

It is possible that a index of similarity that utilizes both types of data
could provide a more accurate estimate of genetic similarity than either
alone as suggested by Cox et al. (1985a) who noted that the biases of these
measures would tend to cancel out each other. There are a number of ways
that these measures could be combined, ranging from simply adding the
CP and SI estimate to combining the multivariate ancestral parent

51

contribution and marker loci data sets and calculating a distance based on
a principal component analysis of the data (Goodman, 1972). It would be
useful to attempt to integrate the two data types in light of some
measurement of actual genetic distance between individuals to see what
measure is most predictive.

List of References

List of References

Apuya. N.R.. B. Frazier, P. Kiem. E..). Roth. anf KG. Lark. 1988. Restriction fragment
length polymorphisms as genetic markers in soybean. Glycine max L. Merrill.. Theor.
Appl. Genet 75:889-901.

Buttery. S.R.. and RI. Buzzell. 1968. Peroxidase activity in seeds of soybean varieties.
Crop Sci. 8:722-725.

Cardy. E..]. and W.D. Beversdorf. 1984. A procedure for starch gel electrophoretic
detection of isozymes of soybean (Glycine max L. Merr.). Departemnt of Crop Science
Technical Bulletin 1 19/8401. University of Guelph, Guelph Ontario. Canada. N1G 2W1.

Chiang, Y.C.. 1986. Genetic and quantitative variation in wild soybean (Glycine sold)
population. Ph.D. dissertation. Univ. of New Hampshire.

Cox. T.S.. Y.T. Kiang. MB. Gorman. and BM. Rodgers. 1985a. Relationship bewteen
coefficient of parentage and genetic similarity indices in the soybean. Crop Sci. 25:529-
532.

Cox. T.S.. G.L. Lookhart. D.E. Walker. L.G. Barrel. LD. Albers. and D.M. Rodgers. 1985b.

Genetic relationships among hard red winter wheat cultivars as evaluated by pedigree
analysis and gliadin polyacrylamide gel electrophoretic patterns. Crop Sci. 25:1058-
1063.

Delanney. x. D.M. Rodgers. and RG. Palmer. 1983. Realtive contribution among
ancestral lines to North American soybean cultivars. Crop Sci. 23:944-949.

Doong. J.Y.I-l., and Y.T. Kiang. 1987. Cultivar identification by isozyme analysis. Soy.
Genet. Newsl. 14:189-225.

Fehr. W.R.. and L.B. Ortiz. 1975. Registration of a soybean germplasm population. Crop
Sci 15:739.

Goodman. M.M.. 1972. Distance analysis in biology. Syst. 2001. 21:174— 186.

Gorrnan. MB. 1983. An electrophoretic analysis of the genetic variation in the wild
and cultivated soybean germplasm. Ph.D. dissertation. Univ. of New Hampshire.

Kempthorne. 0.. 1969. An Introduction to Genetic Statistics. Iowa State University
Press. Ames.

Kiem. P.. R.C. Schumaker. and RG. Palmer. 1989. Restriction fragment length
polymorphism diversity in soybean. Theor. Appl. Genet. 77:786-792.

Lee. M.. EB. Godshalk. KR. Lamkey. and W.W. Woodman. 1989. Association of
restriction fragment length polymorphisms among maize inbreds with agronomic
performance oftheir crosses. Crp Sci. 29:1067-1071.

Malecot. G.. 1948. Les Mathemattques de L'heredite. Masson and Cie. Paris

52

53

Nei. M.. 1972. Genetic distance between populations. Am. Nat. 106:283-292.

Ramshaw. JAM" JA. Coyne. and RC. Lewontin. 1979. The sensitivity of gel
electrophoresis as a dcctector of genetic variation. Genet. 93:1019-1037.

St. Martin. SK. 1982. Efi'ective population size for the soybean improvement program
inmaturity groups 00 to 1V. Crop Sci. 22:151-152.

Soka1.R.R.. and P.H.A. Sneath. 1963. Principles of Numerical Taxonomy. Freeman.
San Francisco.

Appendices

Appendix A. Summary of the frequency of the monitored allele in the 62 elite lines at the 19 marker loci.

Enp Per Sod Fla Pgi Acp Pgd1

Pgml Dial

1'th Idbg ACO; ACO; Mpil

t1 br

"1

Name

 

1.0 0.0 0.0 0.0 1.0 0.0 1.0
1.0 0.0 0.0 0.5 0.0 0.0 1.0

0.0

0.0

1.0
1.0

1.0
1.0
1.0
1.0
1.0
1.0
0.0

0.0
0.0

1.0
1.0
1.0
0.0
0.0
0.0
0.0
0.0
1.0
1.0
0.0

1.0
1.0
1.0
1.0
0.5
1.0
0.0
0.0
1.0

0.0 1.0 0.0 1.0 0.5
0.0 1.0 1.0 1.0 0.5

A80-147002
A82-161035

0.0

1.0

1.0 0.0 0.0 0.0 0.0 0.0 1.0

0.0
0.5
1.0
1.0

1.0
0.5
1.0

0.0
0.5
0.5

0.0
0.0
1.0

A83-271010 0.0 1.0 1.0 1.0 0.5

A84-185032

1.0 1.0 0.0 1.0 0.0 0.0 1.0
1.0 0.0 0.0 0.0 0.0 0.0 1.0

1.0 1.0 1.0 0.0 1.0

1.0

ABS-192034 0.5 0.0 1.0 1.0 1.0

ABS—291010

1.0 0.0 0.0 0.0 0.0 0.0 1.0
1.0 1.0 0.0 0.0 0.0 0.0 1.0

1.0
0.0

0.0

1.0
1.0

1.0 1.0 1.0 1.0 1.0

1.0

ABS-292023 0.0 0.0 1.0 1.0 1.0

1.0 1.0 0.0 1.0 0.0 0.0 1.0
1.0 0.0 0.0 0.0 0.5 0.0 1.0

1.0
1.0

0.5 1.0

1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
0.0

1.0

0.0
0.0

0.5 1.0 1.0 0.0 1.0

A86-102004 1.0 0.0 0.0 1.0 1.0

ABS-293033
A86-103017

1.0

0.0
0.0

1.0 1.0 0.0 0.0 0.0 0.0 1.0

1.0
0.0
1.0
0.0
0.0
1.0
1.0
1.0
1.0
1.0

1.0 0.5

1.0
1.0
0.0

1.0 0.0 1.0 0.0 0.0

1.0 0.0 0.0 1.0 1.0 0.0 0.0

1.0
0.5

0.0

1.0

A86-103027 0.0 0.0 1.0 1.0 1.0

1.0 0.0 0.0 0.0 0.0 0.0 1.0

0.0

1.0

1.0

1.0 0.0 1.0 0.0 1.0
1.0 0.0 1.0 1.0 0.0

A86-202026
A87-198005
A87-296011
A87-29701S

54

1.0 0.0 0.0 0.5 1.0 0.0 0.0
1.0 0.0 0.0 1.0 1.0 0.0 1.0
1.0 0.0 0.0 1.0 0.0 0.0 0.0
1.0 1.0 0.0 1.0 1.0 0.0 0.0

0.0
0.0

1.0

1.0

0.0
1.0

1.0

1.0
1.0
1.0
1.0
1.0
0.0
1.0
1.0
1.0
1.0
1.0

1250 1.0 1.0 1.0 0.0 1.0 0.0

1.0

1.0
0.0

1.0 0.0 1.0 1.0 1.0

1.0

1.0

1.0 0.0 1.0 1.0 1.0

0.0

1.0
0.0
0.0

1.0

1.0
0.0
1.0
0.0
1.0
0.0
1.0
1.0
0.0
0.0
0.0
1.0
1.0
1.0
1.0

0.0

AP 1989 1.0 1.0 1.0 0.0 1.0

1.0 0.0 0.0 0.0 1.0 0.0 1.0
1.0 0.0 0.0 0.0 0.0 0.0 1.0

1.0

0.0

A1937 1.0 0.0 0.0 1.0 1.0

0.0

1.0

A2234 1.0 0.0 0.0 1.0 0.0

1.0 1.0 0.0 0.5 0.5 0.0 1.0

0.0

1.0

1.0
0.0

A2943 1.0 1.0 0.0 1.0 0.0
BSR 101 1.0 1.0 1.0 0.0 1.0

1.0 0.0 0.0 0.0 1.0 0.0 1.0

0.0 0.5

0.0

0.0

1.0 0.5 0.0 1.0 0.0 0.0 1.0
1.0 0.0 0.0 0.0 1.0 0.0 0.0

1.0

0.0

1.0

0.0

1.0 0.0 1.0 1.0 0.5

BURLISON 0.0 0.0 0.0 1.0 1.0

BSR 201

0.0 1.0
0.0 1.0
0.0 1.0
0.0 1.0

0.0 1.0

0.0
0.0
0.0

1.0
1.0
1.0
1.0
1.0
1.0

0.5

1.0 0.0 0.0 0.0 1.0 0.0 1.0

0.0

C1664 0.0 0.0 1.0 1.0 1.0

1.0 1.0 0.0 1.0 0.0 0.0 1.0
1.0 0.0 0.0 1.0 0.0 0.0 0.5
1.0 0.0 0.0 1.0 0.0 0.0 0.0
1.0 0.0 0.0 1.0 1.0 0.0 1.0

0.0
0.0

0.0

0.0

7299 1.0 1.0 1.0 0.0 0.0 0.5

7260 1.0 1.0 1.0 1.0 1.0

0.0

1.0

0.0 1.0
0.0 1.0
0.0

0.0
1.0 0.0

0.0
1.0

8252 1.0 1.0 0.0 0.0 0.0 0.0

1.0 0.0 0.0 0.0 0.0 0.0 1.0

1.0
1.0
1.0

CORSOY 1.0 1.0 1.0 0.0 1.0 1.0

CENTURY 1.0 0.0 0.0 1.0 0.0

1.0 0.0 0.0 0.0 1.0 0.0 1.0

1.0

0.0
0.0

1.0

0.0

CHAPMAN 1.0 0.0 1.0 1.0 0.0

1.0 0.0 0.0 0.5 1.0 0.0 1.0

0.0 0.0

0.0 1.0

1.0
1.0
1.0
1.0
1.0

1.0

CONRAD 1.0 0.0 0.0 1.0 1.0

1.0 1.0 0.0 0.0 1.0 0.0 1.0
1.0 0.0 0.0 1.0 1.0 0.0 1.0

1.0

1.0
0.0
1.0

0.0 1.0

0.0

0.0

0.0
0.0
0.5

1.0
1.0
1.0

DSR-262 1.0 0.0 0.0 1.0 0.0

1.0 1.0 0.0 1.0 0.0 0.0 1.0
1.0 1.0 0.0 0.0 0.0 0.0 1.0

1.0

0.0

884108 1.0 1.0 1.0 0.0 1.0

0.0 0.0 0.0

1.0

884150 1.0 0.0 1.0 1.0 0.0

Appendix A. Continued

Acp Pgdl

Pgml Dial Enp Per Sod Fle Pgr'

Idhg Ac04 Acoz Mp1;

1'th

r

81 br

"1

Name

 

1.0 1.0 0.0 1.0 0.0 0.0 1.0
1.0 1.0 0.0 1.0 1.0 0.0 1.0
1.0 1.0 0.0 0.5 1.0 0.0 1.0

0.0
0.0

0.0

0.5
0.0
0.0
0.0

0.0
0.0
1.0
1.0

1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0

0.0
1.0

0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.0
0.0

1.0
1.0
0.0

0.5
0.5

E84159 0.0 1.0 0.0 1.0 0.5

1.0

884165 0.0 1.0 1.0 1.0 0.5

0.0
0.0

885100 0.5 0.0 0.0 1.0 0.0

1.0 0.0 0.0 0.0 1.0 0.0 0.5
1.0 1.0 0.0 0.0 1.0 0.0 1.0

1.0

E85166 0.0 0.0 0.0 1.0 0.0

E85168? 1.0 0.0 0.5 1.0 0.0

0.0

0.0

0.5

0.0

1.0
1.0
1.0
0.0
0.5

1.0 0.0 0.0 0.5 0.5 0.0 1.0
1.0 0.0 0.0 1.0 1.0 0.0 1.0
1.0 1.0 0.0 0.5 0.5 0.0 1.0
1.0 0.5 0.0 0.0 0.0 0.0 1.0
1.0 0.0 0.0 0.0 1.0 0.0 1.0
1.0 0.0 0.0 0.0 1.0 0.0 0.0
1.0 1.0 0.0 0.5 1.0 0.0 1.0
1.0 0.0 0.0 0.0 0.0 0.0 1.0
1.0 0.0 0.0 0.0 1.0 0.0 1.0
1.0 0.0 0.0 0.1 0.0 0.0 1.0
1.0 0.0 0.0 0.0 0.0 0.0 1.0
1.0 0.5 0.0 0.0 1.0 0.0 1.0
1.0 1.0 0.0 0.0 1.0 0.0 1.0
1.0 0.0 0.0 0.5 0.0 0.0 1.0
1.0 0.0 0.0 0.0 0.0 0.0 1.0
1.0 0.0 0.0 0.0 1.0 0.0 1.0
1.0 0.0 0.0 0.0 1.0 0.0 0.0
1.0 0.0 0.0 0.0 1.0 0.0 0.0
1.0 1.0 0.0 1.0 0.0 0.0 1.0
1.0 1.0 0.0 0.5 0.0 0.0 1.0
1.0 0.0 1.0 0.0 1.0 0.0 1.0

1.0
1.0
0.0
1.0
1.0

0.5

1.0
0.0

1.0
1.0
0.0
0.0
0.0

1.0

886339 1.0 0.0 0.5 1.0 0.0

1.0

886348 1.0 0.0 1.0 1.0 0.0

E87223

0.5

1.0

1.0 1.0 0.5 1.0 0.0

0.0
1.0
1.0

1.0

388080 0.5 0.0 1.0 1.0 0.5

1.0
1.0

1.0
0.0
1.0
0.0
1.0
1.0
1.0
0.0
1.0
0.0
1.0
0.0
1.0
1.0
1.0
1.0
1.0
1.0

ELGIN 1.0 0.0 1.0 1.0 0.0
ZANE 1.0 1.0 1.0 1.0 0.0 1.0

0.0
1.0
0.0

HACK 0.0 1.0 0.0 1.0 0.5

HARDIN 1.0 1.0 1.0 0.0 1.0

1.0

0.0
0.0

1.0

55

1.0

1.0

HW8008 1.0 0.0 0.0 1.0 0.0

1.0 1.0 1.0
1.0

0.0
0.0

1.0

J-231 1.0 1.0 1.0 1.0 0.0

KENWOOD 1.0 0.0 1.0 1.0 0.0
LN82-296 1.0 0.0 0.0 1.0 1.0

1.0
0.0

1.0

1.0
0.0
0.0
1.0 0.0
1.0 1.0
0.0 0.0
1.0 1.0

0.0

1.0
0.0

1.0

1.0
0.0
0.0
1.0
0.0
0.0

1.0

1.0

1.0

1.0

1.0 1.0 1.0 0.0 1.0
1484-916 0.0 1.0 1.0 1.0 0.5

1182-946

1.0

1.0
0.0
0.0
1.0
1.0
0.0

1.0
1.0
1.0 0.0

1.0

100

823-12 0.0 1.0 1.0 0.0 1.0

1.0

1.0
0.0

82596 1.0 0.0 0.0 1.0 1.0

1.0

81884 0.0 1.0 1.0 1.0 0.5
319-90 1.0 0.0 0.0 0.0 0.0

1.0

1.0

1.0

0.0 1.0

1.0
0.0

PELLA 1.0 0.0 0.0 1.0 0.0

1.0 0.0
1.0 0.0
1.0 0.0

0.0 1.0

0.0

9271 1.0 0.0 1.0 1.0 1.0

0.0
1.0

1.0 1.0
1.0

0.0
1.0

1.0 0.0 1.0 1.0 1.0
SIBLEY 0.0 1.0 1.0 0.0 0.5

9292

1.0

1.0 1.0 0.0 0.5 1.0 0.0 1.0
1.0 0.0 0.0 0.0 1.0 0.0 1.0

1.0 0.0
1.0 0.0

1.0

1.0

STURDY 1.0 1.0 1.0 1.0 0.0

1.0

0.0

0.0

frequencies of either 1.0 or 0.0 while other lines were heterogeneous and received a frequency of

reported in Table 7 and were determined as described in the materials and methods. Most lines were

homogenous and therefore had allele

0.5.

The frequencies are for the alleles

56
Appendix B. Coemcient of parentage between the 62 elite lines.

CORSOY CENTURY AP 1989 HW8008 82596 A1937 BURLISON 823-12 SIBLEY

CORSOY 1.000 .211 .570 .211 .141 .250 .176 .141 .426
CENTURY .211 1.000 .387 .719 .201 .154 .594 .188 .219
AP 1989 .570 .387 1.000 .356 .162 .209 .260 .141 .301
HW8008 .211 .719 .356 1.000 .272 .137 .439 .207 .203
82596 .141 .201 .162 .272 1.000 .229 .180 .182 .144
A1937 .250 .154 .209 .137 .229 1.000 .129 .057 .348
BURLISON .176 .594 .260 .439 .180 .129 1.000 .142 .166
823-12 .141 .188 .141 .207 .182 .057 .142 1.000 .128
SIBLEY .426 .219 .301 .203 .144 .348 .166 .128 1.000
A83-271010 .338 .311 .325 .298 .227 .299 .218 .116 .221
BSR 201 .285 .247 .309 .216 .247 .384 .184 .088 .207
7299 .160 .194 .182 .172 .086 .195 .125 .309 .204
J-231 .660 .315 .475 .297 .154 .316 .221 .126 .405
PELLA .070 .406 .263 .379 .244 .298 .256 .099 .129
9271 .297 .150 .218 .155 .211 .289 .127 .087 .183
9292 .297 .150 .218 .155 .211 .289 .127 .087 .183
STURDY .336 .603 .349 .457 .173 .176 .377 .155 .293
A 2934 ..141 .221 .163 .196 .166 .169 .177 .152 .149
A80-147002 .285 .295 .300 .275 .233 .361 .206 .093 .197
A82-161035 .320 .245 .318 .242 .258 .337 .185 .122 .215
A84-185032 .141 .301 .264 .286 .231 .210 .208 .140 .162
ABS-291010 .160 .160 .169 .164 .291 .732 .136 .080 .238
1250 .211 .290 .225 .252 .197 .227 .221 .169 .217
CONRAD .088 .252 .179 .248 .193 .200 .177 .092 .110
C1664 .141 .561 .243 .440 .235 .198 .349 .143 .156
DSR-262 .143 .157 .177 .144 .202 .304 .129 .063 .124
E84150 .106 .538 .218 .398 .181 .188 .336 .114 .130
884159 .250 .121 .188 .125 .192 .280 .109 .070 .148
E84165 .250 .121 .188 .125 .192 .280 .109 .070 .148
HACK .035 .287 .187 .260 .288 .429 .198 .068 .121
LN82-296 .141 .703 .325 .549 .223 .226 .425 .143 .174
M82-946 .292 .262 .267 .271 .230 .301 .190 .144 .378
81884 .469 .173 .316 .161 .210 .397 .161 .082 .249
819-90 .035 .219 .139 .201 .124 .153 .136 .304 .077
ZANE .285 .287 .294 .276 .233 .319 .199 .099 .192
ABS-192034 .178 .369 .229 .283 .248 .679 .248 .084 .257
ABS-292023 .088 .200 .149 .192 .309 .425 .163 .105 .140
A86-102004 .285 .158 .221 .171 .254 .626 .132 .099 .272
A86-103017 .266 .226 .229 .228 .239 .239 .178 .155 .207
A86-103027 .143 .221 .198 .198 .258 .715 .164 .062 .234
A86-202026 .266 .226 .229 .228 .239 .239 .178 .155 .207
A87-198005 .285 .295 .300 .275 .233 .361 .206 .093 .197
A87-296011 .160 .234 .211 .223 .293 .425 .174 .082 .161
A87-297015 .143 .163 .158 .149 .119 .184 .111 .300 .111
BSR 101 .141 .402 .301 .519 .300 .242 .263 .158 .166
7260 .293 .404 .353 .358 .185 .303 .269 .131 .302
8252 .141 .348 .244 .315 .221 .263 .239 .134 .173
E85100 .143 .179 .170 .171 .193 .288 .130 .061 .119
E85166 .283 .398 .314 .341 .226 .251 .290 .164 .229
885168 .283 .398 .314 .341 .226 .251 .290 .164 .229
886348 .213 .324 .273 .294 .227 .308 .224 .116 .179
HARDIN .938 .211 .542 .208 .163 .259 .177 .135 .405
M84-916 .356 .192 .265 .178 .229 .307 .157 .096 .271
CHAPMAN .455 .176 .309 .182 .245 .375 .148 .103 .246
888080 .248 .629 .341 .475 .203 .317 .388 .136 .293
ELGIN .169 .216 .186 .202 .234 .270 .169 .108 .169
ABS-293032 .137 .400 .217 .316 .250 .313 .268 .110 .167
E84098 .469 .143 .296 .143 .162 .240 .127 .088 .223
E86339 .155 .285 .217 .258 .227 .262 .206 .123 .164
XENWOOD .210 .185 .198 .169 .231 .635 .149 .082 .258
A2234 .013 .161 .098 .164 .303 .411 .127 .067 .090
ABS-293033 .356 .237 .281 .211 .210 .325 .188 .128 .241

E87223 .585 .214 .378 .207 .187 .260 .172 .124 .297

5 7
Appendix B. continued

A83-271010 BSR 201 7299 J-231 PELLA 9271 9292 STURDY A 2934
CORSOY .338 .285 .160 .660 .070 .297 .297 .336 .141
CENTURY .311 .247 .194 .315 .406 .150 .150 .603 .221
AP 1989 .325 .309 .182 .475 .263 .218 .218 .349 .163
HW8008 .298 .216 .172 .297 .379 .155 .155 .457 .196
82596 .227 .247 .086 .154 .244 .211 .211 .173 .166
A1937 .299 .384 .195 .316 .298 .289 .289 .176 .169
BURLISON .218 .184 .125 .221 .256 .127 .127 .377 .177
823-12 .116 .088 .309 .126 .099 .087 .087 .155 .152
SIBLEY .221 .207 .204 .405 .129 .183 .183 .293 .149
A83-271010 1.000 .317 .145 .324 .321 .244 .244 .269 .175
BSR 201 .317 1.000 .140 .275 .327 .281 .281 .229 .188
7299 .145 .140 1.000 .210 .159 .093 .093 .158 .100
J-231 .324 .275 .210 1.000 .215 .239 .239 .330 .146
PELLA .321 .327 .159 .215 1.000 .247 .247 .263 .197
9271 .244 .281 .093 .239 .247 1.000 .500 .170 .199
9292 .244 .281 .093 .239 .247 .500 1.000 .170 .199
STURDY .269 .229 .158 .330 .263 .170 .170 1.000 .182
A 2934 .175 .188 .100 .146 .197 .199 .199 .182 1.000
A80-147002 .458 .362 .146 .298 .632 .286 .286 .248 .188
A82-161035 .304 .499 .131 .285 .287 .264 .264 .232 .183
A84-185032 .242 .290 .148 .186 .277 .173 .173 .227 .174
A85-291010 .307 .396 .145 .221 .336 .299 .299 .161 .186
1250 .236 .248 .138 .207 .224 .192 .192 .248 .286
CONRAD .262 .218 .105 .155 .302 .209 .209 .179 .342
C1664 .249 .236 .129 .206 .313 .167 .167 .355 .176
DSR-262 .229 .599 .096 .158 .277 .268 .268 .141 .351
884150 .226 .223 .125 .179 .318 .202 .202 .330 .200
884159 .222 .267 .089 .206 .241 .289 .289 .138 .176
884165 .222 .267 .089 .206 .241 .289 .289 .138 .176
HACK .331 .420 .136 .161 .481 .331 .331 .201 .223
LN82-296 .316 .287 .177 .265 .457 .235 .235 .433 .227
M82-946 .328 .239 .163 .308 .232 .199 .199 .263 .164
81884 .559 .373 .125 .356 .248 .306 .306 .218 .171
819-90 .166 .171 .334 .112 .510 .126 .126 .144 .123
ZANE .312 .331 .140 .292 .626 .323 .323 .241 .204
A85-192034 .313 .382 .174 .263 .364 .268 .268 .274 .193
ABS-292023 .297 .393 .105 .143 .356 .284 .284 .169 .212
A86-102004 .271 .320 .148 .285 .257 .286 .286 .180 .173
A86-103017 .239 .252 .120 .230 .220 .238 .238 .216 .232
A86-103027 .315 .402 .166 .238 .389 .310 .310 .189 .196
A86-202026 .239 .252 .120 .230 .220 .238 .238 .216 .232
A87-198005 .396 .424 .146 .298 .527 .286 .286 .248 .188
A87-296011 .326 .471 .127 .209 .415 .302 .302 .198 .195
A87-297015 .235 .188 .327 .153 .326 .145 .145 .136 .119
BSR 101 .306 .304 .155 .239 .404 .200 .200 .280 .185
7260 .292 .278 .349 .360 .325 .186 .186 .313 .177
8252 .278 .288 .148 .211 .612 .219 .219 .256 .241
885100 .296 .277 .094 .168 .425- .268 .268 .150 .165
885166 .352 .292 .149 .288 .431 .223 .223 .312 .223
885168 .352 .292 .149 .288 .431 .223 .223 .312 .223
886348 .302 .325 .142 .256 .621 .248 .248 .252 .200
HARDIN .333 .292 .157 .625 .089 .294 .294 .324 .143
M84-916 .265 .364 .118 .290 .218 .243 .243 .293 .160
CHAPMAN .318 .467 .122 .347 .262 .350 .350 .218 .195
888080 .298 .294 .208 .340 .351 .191 .191 .420 .190
EIGIN .235 .274 .105 .181 .258 .218 .218 .190 .177
ABS-293032 .280 .327 .129 .196 .344 .232 .232 .281 .197
884098 .237 .246 .106 .334 .159 .273 .273 .191 .160
886339 .253 .281 .121 .197 .435 .214 .214 .223 .195
KENWOOD .267 .329 .150 .249 .278 .253 .253 .183 .173
A2234 .269 .373 .089 .092 .431 .478 .478 .124 .276
ABS-293033 .291 .448 .135 .293 .244 .259 .259 .241 .233

887223 .287 .280 .133 .421 .164 .257 .257 .263 .159

58
Appendix B. continued.

A80- A82- A84- A85-
147002 161035 185032 291010 1250 CONRAD C1664 BSR-262 884150

CORSOY .285 .320 .141 .160 .211 .088 .141 .143 .106
CENTURY .295 .245 .301 .160 .290 .252 .561 .157 .538
AP 1989 .300 .318 .264 .169 .225 .179 .243 .177 .218
HW8008 .275 .242 .286 .164 .252 .248 .440 .144 .398
82596 .233 .258 .231 .291 .197 .193 .235 .202 .181
A1937 .361 .337 .210 .732 .227 .200 .198 .304 .188
BURLISON .206 .185 .208 .136 .221 .177 .349 .129 .336
823-12 .093 .122 .140 .080 .169 .092 .143 .063 .114
SIBLEY .197 .215 .162 .238 .217 .110 .156 .124 .130
A83-271010 .458 .304 .242 .307 .236 .262 .249 .229 .226
BSR 201 .362 .499 .290 .396 .248 .218 .236 .599 .223
7299 .146 .131 .148 .145 .138 .105 .129 .096 .125
J-231 .298 .285 .186 .221 .207 .155 .206 .158 .179
PELLA .632 .287 .277 .336 .224 .302 .313 .277 .318
9271 .286 .264 .173 .299 .192 .209 .167 .268 .202
9292 .286 .264 .173 .299 .192 .209 .167 .268 .202
STURDY .248 .232 .227 .161 .248 .179 .355 .141 .330
A 2934 .188 .183 .174 .186 .286 .342 .176 .351 .200
ABC-147002 1.000 .330 .244 .376 .237 .246 .257 .284 .251
A82-161035 .330 1.000 .416 .357 .244 .202 .233 .334 .207
A84-185032 .244 .416 1.000 .238 .228 .197 .237 .206 .214
ABS-291010 .376 .357 .238 1.000 .245 .228 .226 .328 .208
1250 .237 .244 .228 .245 1.000 .166 .222 .179 .200
CONRAD .246 .202 .197 .228 .166 1.000 .208 .383 .232
C1664 .257 .233 .237 .226 .222 .208 1.000 .177 .341
BSR-262 .284 .334 .206 .328 .179 .383 .177 1.000 .210
884150 .251 .207 .214 .208 .200 .232 .341 .210 1.000
884159 .273 .245 .156 .292 .159 .196 .151 .260 .356
884165 .273 .245 .156 .292 .159 .196 .151 .260 .356
HACK .420 .343 .280 .479 .248 .314 .287 .373 .307
LN82-296 .341 .266 .289 .248 .257 .295 .437 .235 .446
M82-946 .246 .248 .212 .272 .223 .194 .214 .169 .180
81884 .593 .351 .198 .392 .235 .180 .190 .276 .174
819-90 .323 .150 .155 .172 .148 .161 .165 .147 .173
ZANE .486 .304 .231 .332 .217 .259 .246 .286 .264
ABS-192034 .368 .329 .251 .562 .254 .235 .304 .300 .293
ABS-292023 .366 .344 .256 .575 .259 .239 .247 .327 .229
A86-102004 .309 .301 .199 .503 .213 .192 .191 .262 .180
A86-103017 .247 .255 .208 .260 .391 .175 .204 .199 .186
A86-103027 .391 .340 .245 .606 .237 .257 .242 .339 .247
A86-202026 .247 .255 .208 .260 .391 .175 .204 .199 .186
A87-198005 .517 .393 .244 .376 .237 .246 .257 .315 .251
A87-296011 .405 .413 .259 .465 .247 .257 .256 .365 .245
A87-297015 .323 .171 .139 .192 .155 .132 .137 .151 .139
BSR 101 .320 .299 .304 .284 .234 .262 .311 .226 .277
7260 .293 .265 .289 .255 .235 .208 .270 .185 .251
8252 .434 .266 .252 .291 .404 .234 .268 .228 .259
885100 .414 .246 .179 .312 .170 .209 .183 .254 .204
885166 .422 .283 .260 .260 .288 .209 .282 .209 .263
885168 .422 .283 .260 .260 .288 .209 .282 .209 .263
886348 .468 .293 .243 .326 .233 .237 .262 .254 .255
HARDIN .288 .318 .147 .177 .211 .097 .147 .154 .114
M84-916 .283 .342 .189 .300 .218 .159 .182 .250 .164
CHAPMAN .354 .470 .220 .378 .231 .205 .197 .344 .197
888080 .305 .273 .269 .277 .250 .223 .389 .201 .371
EIBIN .262 .249 .205 .292 .228 .182 .210 .217 .188
ABS-293032 .319 .285 .249 .342 .254 .226 .310 .256 .293
884098 .247 .244 .137 .216 .160 .154 .134 .208 .326
886339 .345 .252 .224 .284 .229 .205 .238 .221 .224
KENWOOD .311 .293 .207 .512 .228 .191 .204 .261 .188
A2234 .387 .316 .231 .474 .211 .339 .231 .428 .320
ABS-293033 .309 .434 .219 .331 .417 .177 .212 .298 .192

887223 .274 .284 .173 .226 .220 .135 .175 .180 .147

59
Appendix B. continued.

884159 884165 HACK LN82-296 M82-946 81884 819—90 ZANE .A85-192034

CORSOY .250 .250 .035 .141 .292 .469 .035 .285 .178
CENTURY .121 .121 .287 .703 .262 .173 .219 .287 .369
AP 1989 .188 .188 .187 .325 .267 .316 .139 .294 .229
HW8008 .125 .125 .260 .549 .271 .161 .201 .276 .283
82596 .192 .192 .288 .223 .230 .210 .124 .233 .248
A1937 .280 .280 .429 .226 .301 .397 .153 .319 .679
BURLISON .109 .109 .198 .425 .190 .161 .136 .199 .248
823-12 .070 .070 .068 .143 .144 .082 .304 .099 .084
SIBLEY .148 .148 .121 .174 .378 .249 .077 .192 .257
A83-271010 .222 .222 .331 .316 .328 .559 .166 .312 .313
BSR 201 .267 .267 .420 .287 .239 .373 .171 .331 .382
7299 .089 .089 .136 .177 .163 .125 .334 .140 .174
J-231 .206 .206 .161 .265 .308 .356 .112 .292 .263
PELLA .241 .241 .481 .457 .232 .248 .510 .626 .364
9271 .289 .289 .331 .235 .199 .306 .126 .323 .268
9292 .289 .289 .331 .235 .199 .306 .126 .323 .268
STURDY .138 .138 .201 .433 .263 .218 .144 .241 .274
A 2934 .176 .176 .223 .227 .164 .171 .123 .204 .193
ABC-147002 .273 .273 .420 .341 .246 .593 .323 .486 .368
A82-161035 .245 .245 .343 .266 .248 .351 .150 .304 .329
A84-185032 .156 .156 .280 .289 .212 .198 .155 .231 .251
ABS-291010 .292 .292 .479 .248 .272 .392 .172 .332 .562
1250 .159 .159 .248 .257 .223 .235 .148 .217 .254
CONRAD .196 .196 .314 .295 .194 .180 .161 .259 .235
C1664 .151 .151 .287 .437 .214 .190 .165 .246 .304
BSR-262 .260 .260 .373 .235 .169 .276 .147 .286 .300
884150 .356 .356 .307 .446 .180 .174 .173 .264 .293
884159 1.000 .590 .312 .199 .170 .288 .128 .290 .257
884165 .590 1.000 .312 .199 .170 .288 .128 .290 .257
HACK .312 .312 1.000 .402 .229 .339 .249 .407 .457
LN82-296 .199 .199 .402 1.000 .247 .211 .241 .352 .367
M82-946 .170 .170 .229 .247 1.000 .245 .125 .237 .273
81884 .288 .288 .339 .211 .245 1.000 .127 .324 .349
819-90 .128 .128 .249 .241 .125 .127 1.000 .320 .190
ZANE .290 .290 .407 .352 .237 .324 .320 1.000 .331
ABS-192034 .257 .257 .457 .367 .273 .349 .190 .331 1.000
ABS-292023 .276 .276 .497 .278 .227 .354 .184 .325 .427
A86-102004 .261 .261 .347 .217 .262 .339 .135 .361 .457
A86-103017 .201 .201 .257 .232 .223 .258 .133 .310 .245
A86-103027 .296 .296 .715 .314 .265 .368 .201 .363 .568
A86-202026 .201 .201 .257 .232 .223 .258 .133 .310 .245
A87-l98005 .273 .273 .420 .341 .246 .476 .270 .433 .368
A87-296011 .293 .293 .509 .315 .239 .371 .214 .372 .427
A87-297015 .144 .144 .219 .183 .132 .299 .416 .249 .192
BSR 101 .182 .182 .371 .403 .257 .222 .211 .308 .306
7260 .165 .165 .279 .365 .278 .244 .173 .279 .311
8252 .200 .200 .364 .357 .228 .241 .329 .421 .309
885100 .244 .244 .369 .258 .171 .380 .217 .360 .288
885166 .195 .195 .297 .353 .248 .388 .227 .346 .296
885168 .195 .195 .297 .353 .248 .388 .227 .346 .296
886348 .235 .235 .391 .350 .230 .295 .319 .452 .337
HARDIN .250 .250 .065 .150 .285 .457 .045 .286 .192
M84-916 .222 .222 .278 .205 .243 .327 .115 .264 .286
CHAPMAN .313 .313 .351 .235 .252 .419 .134 .461 .333
888080 .170 .170 .315 .490 .283 .242 .187 .289 .385
ELGIN .199 .199 .312 .237 .202 .250 .139 .248 .280
A85-293032 .216 .216 .399 .372 .223 .276 .183 .295 .479
884098 .420 .420 .195 .169 .192 .317 .086 .264 .204
886339 .198 .198 .337 .298 .208 .241 .227 .334 .293
XENWOOD .240 .240 .371 .232 .251 .324 .146 .284 .479
A2234 .411 .411 .610 .365 .195 .322 .221 .455 .408
A85-293033 .232 .232 .304 .240 .242 .352 .142 .281 .313

887223 .225 .225 .174 .189 .247 .359 .087 .267 .229

60
Appendix B. continued.

A85- A86- A86- A86- A86- A87- A87- A87-
292023 102004 103017 103027 202026 198005 296011 297015 BSR 101

CORSOY .088 .285 .266 .143 .266 .285 .160 .143 .141
CENTURY .200 .158 .226 .221 .226 .295 .234 .163 .402
AP 1989 .149 .221 .229 .198 .229 .300 .211 .158 .301
HW8008 .192 .171 .228 .198 .228 .275 .223 .149 .519
82596 .309 .254 .239 .258 .239 .233 .293 .119 .300
A1937 .425 .626 .239 .715 .239 .361 .425 .184 .242
BURLISON .163 .132 .178 .164 .178 .206 .174 .111 .263
823-12 .105 .099 .155 .062 .155 .093 .082 .300 .158
SIBLEY .140 .272 .207 .234 .207 .197 .161 .111 .166
A83-271010 .297 .271 .239 .315 .239 .396 .326 .235 .306
BSR 201 .393 .320 .252 .402 .252 .424 .471 .188 .304
7299 .105 .148 .120 .166 .120 .146 .127 .327 .155
J-231 .143 .285 .230 .238 .230 .298 .209 .153 .239
PELLA .356 .257 .220 .389 .220 .527 .415 .326 .404
9271 .284 .286 .238 .310 .238 .286 .302 .145 .200
9292 .284 .286 .238 .310 .238 .286 .302 .145 .200
STURDY .169 .180 .216 .189 .216 .248 .198 .136 .280
A 2934 .212 .173 .232 .196 .232 .188 .195 .119 .185
ABC-147002 .366 .309 .247 .391 .247 .517 .405 .323 .320
A82-161035 .344 .301 .255 .340 .255 .393 .413 .171 .299
A84-185032 .256 .199 .208 .245 .208 .244 .259 .139 .304
A85-291010 .575 .503 .260 .606 .260 .376 .465 .192 .284
1250 .259 .213 .391 .237 .391 .237 .247 .155 .234
CONRAD .239 .192 .175 .257 .175 .246 .257 .132 .262
C1664 .247 .191 .204 .242 .204 .257 .256 .137 .311
DSR-262 .327 .262 .199 .339 .199 .315 .365 .151 .226
884150 .229 .180 .186 .247 .186 .251 .245 .139 .277
884159 .276 .261 .201 .296 .201 .273 .293 .144 .182
884165 .276 .261 .201 .296 .201 .273 .293 .144 .182
HACK .497 .347 .257 .715 .257 .420 .509 .219 .371
LN82-296 .278 .217 .232 .314 .232 .341 .315 .183 .403
M82-946 .227 .262 .223 .265 .223 .246 .239 .132 .257
81884 .354 .339 .258 .368 .258 .476 .371 .299 .222
819-90 .184 .135 .133 .201 .133 .270 .214 .416 .211
ZANE .325 .361 .310 .363 .310 .433 .372 .249 .308
A85-l92034 .427 .457 .245 .568 .245 .368 .427 .192 .306
ABS-292023 1.000 .343 .260 .461 .260 .366 .467 .189 .304
A86-102004 .343 1.000 .419 .487 .419 .309 .348 .161 .236
A86-103017 .260 .419 1.000 .248 .599 .247 .259 .146 .232
A86-103027 .461 .487 .248 1.000 .248 .391 .467 .202 .306
A86-202026 .260 .419 .599 .248 1.000 .247 .259 .146 .232
A87-l98005 .366 .309 .247 .391 .247 1.000 .555 .357 .320
A87-296011 .467 .348 .259 .467 .259 .555 1.000 .209 .327
A87-297015 .189 .161 .146 .202 .146 .357 .209 1.000 .170
BSR 101 .304 .236 .232 .306 .232 .320 .327 .170 1.000
7260 .222 .246 .212 .291 .212 .293 .258 .157 .317
8252 .307 .235 .306 .313 .306 .382 .331 .241 .319
885100 .308 .253 .195 .328 .195 .449 .327 .304 .231
885166 .278 .234 .253 .274 .253 .457 .289 .314 .300
885168 .278 .234 .253 .274 .253 .457 .289 .314 .300
886348 .378 .267 .230 .349 .230 .415 .366 .242 .317
HARDIN .111 .286 .262 .162 .262 .288 .179 .144 .150
M84-916 .283 .271 .226 .292 .226 .314 .328 .148 .210
CHAPMAN .345 .402 .330 .363 .330 .409 .419 .180 .237
888080 .252 .254 .220 .316 .220 .305 .283 .163 .398
ELGIN .304 .243 .222 .291 .222 .262 .302 .141 .236
ABS-293032 .367 .266 .236 .356 .236 .319 .365 .170 .303
884098 .184 .244 .204 .218 .204 .247 .218 .130 .151
886339 .347 .234 .217 .300 .217 .319 .314 .183 .275
KENWOOD .364 .434 .231 .503 .231 .311 .364 .163 .239
A2234 .481 .380 .281 .510 .281 .387 .482 .199 .292
ABS-293033 .318 .287 .437 .315 .437 .372 .383 .174 .235

887223 .196 .264 .244 .217 .244 .274 .231 .142 .188

6 1
Appendix B. continued.

7260 8252 885100 885166 885168 886348 HARDIN M84-916 CHAPMAN
CORSOY .293 .141 .143 .283 .283 .213 .938 .356 .455
CENTURY .404 .348 .179 .398 .398 .324 .211 .192 .176
AP 1989 .353 .244 .170 .314 .314 .273 .542 .265 .309
HW8008 .358 .315 .171 .341 .341 .294 .208 .178 .182
82596 .185 .221 .193 .226 .226 .227 .163 .229 .245
A1937 .303 .263 .288 .251 .251 .308 .259 .307 .375
BURLISON .269 .239 .130 .290 .290 .224 .177 .157 .148
823-12 .131 .134 .061 .164 .164 .116 .135 .096 .103
SIBLEY .302 .173 .119 .229 .229 .179 .405 .271 .246
A83-271010 .292 .278 .296 .352 .352 .302 .333 .265 .318
BSR 201 .278 .288 .277 .292 .292 .325 .292 .364 .467
7299 .349 .148 .094 .149 .149 .142 .157 .118 .122
J-231 .360 .211 .168 .288 .288 .256 .625 .290 .347
PELLA .325 .612 .425 .431 .431 .621 .089 .218 .262
9271 .186 .219 .268 .223 .223 .248 .294 .243 .350
9292 .186 .219 .268 .223 .223 .248 .294 .243 .350
STURDY .313 .256 .150 .312 .312 .252 .324 .293 .218
A 2934 .177 .241 .165 .223 .223 .200 .143 .160 .195
A80-147002 .293 .434 .414 .422 .422 .468 .288 .283 .354
A82-161035 .265 .266 .246 .283 .283 .293 .318 .342 .470
A84-185032 .289 .252 .179 .260 .260 .243 .147 .189 .220
ABS-291010 .255 .291 .312 .260 .260 .326 .177 .300 .378
1250 .235 .404 .170 .288 .288 .233 .211 .218 .231
CONRAD .208 .234 .209 .209 .209 .237 .097 .159 .205
C1664 .270 .268 .183 .282 .282 .262 .147 .182 .197
DSR-262 .185 .228 .254 .209 .209 .254 .154 .250 .344
884150 .251 .259 .204 .263 .263 .255 .114 .164 .197
884159 .165 .200 .244 .195 .195 .235 .250 .222 .313
884165 .165 .200 .244 .195 .195 .235 .250 .222 .313
HACK .279 .364 .369 .297 .297 .391 .065 .278 .351
LN82-296 .365 .357 .258 .353 .353 .350 .150 .205 .235
M82-946 .278 .228 .171 .248 .248 .230 .285 .243 .252
81884 .244 .241 .380 .388 .388 .295 .457 .327 .419
819-90 .173 .329 .217 .227 .227 .319 .045 .115 .134
ZANE .279 .421 .360 .346 .346 .452 .286 .264 .461
ABS-192034 .311 .309 .288 .296 .296 .337 .192 .286 .333
ABS-292023 .222 .307 .308 .278 .278 .378 .111 .283 .345
A86-102004 .246 .235 .253 .234 .234 .267 .286 .271 .402
A86-103017 .212 .306 .195 .253 .253 .230 .262 .226 .330
A86-103027 .291 .313 .328 .274 .274 .349 .162 .292 .363
A86-202026 .212 .306 .195 .253 .253 .230 .262 .226 .330
A87-198005 .293 .382 .449 .457 .457 .415 .288 .314 .409
A87-29601l .258 .331 .327 .289 .289 .366 .179 .328 .419
A87-297015 .157 .241 .304 .314 .314 .242 .144 .148 .180
BSR 101 .317 .319 .231 .300 .300 .317 .150 .210 .237
7260 1.000 .280 .190 .336 .336 .288 .289 .225 .239
8252 .280 1.000 .298 .360 .360 .427 .150 .218 .247
885100 .190 .298 1.000 .363 .363 .322 .152 .207 .283
885166 .336 .360 .363 1.000 .606 .360 .282 .246 .271
885168 .336 .360 .363 .606 1.000 .360 .282 .246 .271
886348 .288 .427 .322 .360 .360 1.000 .220 .249 .296
HARDIN .289 .150 .152 .282 .282 .220 1.000 .352 .444
M84-916 .225 .218 .207 .246 .246 .249 .352 1.000 .373
CHAPMAN .239 .247 .283 .271 .271 .296 .444 .373 1.000
888080 .378 .301 .204 .330 .330 .305 .249 .228 .238
EIGIN .206 .243 .208 .238 .238 .251 .178 .223 .254
A85-293032 .262 .299 .248 .290 .290 .308 .152 .244 .273
884098 .193 .160 .191 .205 .205 .203 .508 .244 .331
886339 .244 .332 .250 .299 .299 .483 .164 .218 .246
KENWOOD .255 .253 .248 .244 .244 .279 .218 .265 .315
A2234 .185 .321 .430 .246 .246 .350 .041 .257 .410
A85-293033 .247 .330 .225 .289 .289 .274 .349 .345 .448

887223 .249 .192 .175 .261 .261 .232 .558 .289 .355

62
Appendix B. continued.

A85- A85-
888080 ELGIN 293032 884098 886339 KENWOOD A2234 293033 887223
CORSOY .248 .169 .137 .469 .155 .210 .013 .356 .585
CM‘URY .629 .216 .400 .143 .285 .185 .161 .237 .214
AP 1989 .341 .186 .217 .296 .217 .198 .098 .281 .378
HW8008 .475 .202 .316 .143 .258 .169 .164 .211 .207
82596 .203 .234 .250 .162 .227 .231 .303 .210 .187
A1937 .317 .270 .313 .240 .262 .635 .411 .325 .260
BURLISON .388 .169 .268 .127 .206 .149 .127 .188 .172
823-12 .136 .108 .110 .088 .123 .082 .067 .128 .124
SIBLEY .293 .169 .167 .223 .164 .258 .090 .241 .297
A83-271010 .298 .235 .280 .237 .253 .267 .269 .291 .287
BSR 201 .294 .274 .327 .246 .281 .329 .373 .448 .280
7299 .208 .105 .129 .106 .121 .150 .089 .135 .133
J-231 .340 .181 .196 .334 .197 .249 .092 .293 .421
PELLA .351 .258 .344 .159 .435 .278 .431 .244 .164
9271 .191 .218 .232 .273 .214 .253 .478 .259 .257
9292 .191 .218 .232 .273 .214 .253 .478 .259 .257
STURDY .420 .190 .281 .191 .223 .183 .124 .241 .263
A 2934 .190 .177 .197 .160 .195 .173 .276 .233 .159
A80-147002 .305 .262 .319 .247 .345 .311 .387 .309 .274
A82-161035 .273 .249 .285 .244 .252 .293 .316 .434 .284
A84-185032 .269 .205 .249 .137 .224 .207 .231 .219 .173
ABS-291010 .277 .292 .342 .216 .284 .512 .474 .331 .226
1250 .250 .228 .254 .160 .229 .228 .211 .417 .220
CONRAD .223 .182 .226 .154 .205 .191 .339 .177 .135
C1664 .389 .210 .310 .134 .238 .204 .231 .212 .175
DSR-262 .201 .217 .256 .208 .221 .261 .428 .298 .180
884150 .371 .188 .293 .326 .224 .188 .320 .192 .147
884159 .170 .199 .216 .420 .198 .240 .411 .232 .225
884165 .170 .199 .216 .420 .198 .240 .411 .232 .225
HACK .315 .312 .399 .195 .337 .371 .610 .304 .174
LN82-296 .490 .237 .372 .169 .298 .232 .365 .240 .189
M82-946 .283 .202 .223 .192 .208 .251 .195 .242 .247
81884 .242 .250 .276 .317 .241 .324 .322 .352 .359
819-90 .187 .139 .183 .086 .227 .146 .221 .142 .087
ZANE .289 .248 .295 .264 .334 .284 .455 .281 .267
ABS-192034 .385 .280 .479 .204 .293 .479 .408 .313 .229
ABS-292023 .252 .304 .367 .184 .347 .364 .481 .318 .196
A86-102004 .254 .243 .266 .244 .234 .434 .380 .287 .264
A86-103017 .220 .222 .236 .204 .217 .231 .281 .437 .244
A86-103027 .316 .291 .356 .218 .300 .503 .510 .315 .217
A86-202026 .220 .222 .236 .204 .217 .231 .281 .437 .244
A87-198005 .305 .262 .319 .247 .319 .311 .387 .372 .274
A87-296011 .283 .302 .365 .218 .314 .364 .482 .383 .231
A87-297015 .163 .141 .170 .130 .183 .163 .199 .174 .142
BSR 101 .398 .236 .303 .151 .275 .239 .292 .235 .188
7260 .378 .206 .262 .193 .244 .255 .185 .247 .249
8252 .301 .243 .299 .160 .332 .253 .321 .330 .192
885100 .204 .208 .248 .191 .250 .248 .430 .225 .175
885166 .330 .238 .290 .205 .299 .244 .246 .289 .261
885168 .330 .238 .290 .205 .299 .244 .246 .289 .261
886348 .305 .251 .308 .203 .483 .279 .350 .274 .232
HARDIN .249 .178 .152 .508 .164 .218 .041 .349 .558
M84-916 .228 .223 .244 .244 .218 .265 .257 .345 .289
CHAPMAN .238 .254 .273 .331 .246 .315 .410 .448 .355
888080 1.000 .223 .338 .181 .264 .270 .222 .254 .236
EIGIN .223 1.000 .646 .169 .620 .635 .301 .247 .203
A85-293032 .338 .646 1.000 .169 .472 .479 .354 .274 .200
884098 .181 .169 .169 1.000 .163 .204 .261 .248 .319
886339 .264 .620 .472 .163 1.000 .441 .307 .242 .196
KENWOOD .270 .635 .479 .204 .441 1.000 .356 .286 .231
A2234 .222 .301 .354 .261 .307 .356 1.000 .276 .157
ABS-293033 .254 .247 .274 .248 .242 .286 .276 1.000 .301

887223 .236 .203 .200 .319 .196 .231 .157 .301 1.000

63
Appendix C. Similarity Index (SI) between the 62 elite lines

CORSOY CENHURY AP 1989 HW8008 82596 A1937 BURLISON 823-12 SIBLEY

CORSOY 1.000 .579 .842 .526 .632 .632 .553 .553 .658
CENTURY .579 1.000 .526 .947 .842 .737 .763 .553 .500
AP 1989 .842 .526 1.000 .474 .474 .474 .605 .553 .605
HW8008 .526 .947 .474 1.000 .790 .684 .711 .605 .447
82596 .632 .842 .474 .790 1.000 .895 .711 .605 .500
A1937 .632 .737 .474 .684 .895 1.000 .763 .500 .500
BURLISON .553 .763 .605 .711 .711 .763 .974 .526 .526
823-12 .553 .553 .553 .605 .605 .500 .526 .974 .737
SIBLEY .658 .500 .605 .447 .500 .500 .526 .737 .974
A83-271010 .500 .658 .447 .605 .658 .658 .632 .790 .711
BSR 201 .632 .737 .579 .684 .737 .737 .605 .500 .368
7299 .658 .658 .711 .658 .605 .500 .526 .632 .474
J-231 .790 .684 .632 .632 .737 .737 .553 .553 .711
PELLA .526 .842 .579 .790 .684 .790 .868 .395 .447
9271 .579 .579 .526 .526 .737 .737 .500 .605 .447
9292 .658 .658 .553 .605 .816 .711 .526 .658 .526
STURDY .763 .763 .763 .711 .605 .605 .632 .553 .579
A2943 -.737 .684 .632 .632 .632 .579 .500 .368 .447
A80-l47002 .605 .658 .553 .605 .553 .658 .737 .579 .711
A82-161035 .526 .579 .526 .526 .579 .579 .553 .737 .737
A84-185032 .763 .500 .711 .447 .605 .605 .421 .632 .526
ABS-291010 .737 .737 .579 .684 .895 .790 .605 .711 .605
1250 .737 .526 .684 .579 .579 .579 .447 .658 .395
CONRAD .605 .816 .605 .763 .763 .763 .790 .605 .579
C1664 .632 .737 .579 .684 .684 .790 .868 .605 .605
DSR-262 .579 .790 .526 .737 .737 .842 .711 .395 .395
884150 .658 .816 .553 .763 .711 .605 .579 .579 .526
884159 .526 .579 .474 .526 .632 .632 .553 .605 .474
884165 .711 .553 .658 .500 .553 .553 .579 .632 .605
HACK .526 .579 .474 .632 .684 .684 .711 .658 .579
LN82-296 .632 .947 .579 .895 .895 .790 .816 .605 .500
M82-946 .763 .395 .605 .447 .553 .553 .368 .684 .684
81884 .658 .500 .500 .553 .500 .605 .579 .632 .658
819-90 .632 .737 .684 .684 .684 .684 .711 .500 .658
ZANE .684 .790 .632 .737 .632 .737 .711 .553 .605
ABS-192034 .658 .711 .500 .711 .868 .763 .632 .684 .579
ABS-292023 .684 .579 .526 .632 .632 .526 .553 .553 .447
A86-102004 .553 .816 .500 .763 .868 .921 .790 .579 .526
A86-103017 .711 .816 .658 .763 .711 .605 .579 .632 .579
A86-103027 .526 .526 .579 .474 .684 .684 .711 .658 .605
A86-202026 .658 .763 .605 .816 .763 .658 .632 .790 .579
A87-198005 .658 .658 .658 .605 .605 .605 .632 .447 i .579
A87-296011 .684 .684 .737 .632 .632 .632 .658 .605 .658
A87-297015 .526 .632 .684 .579 .684 .684 .658 .553 .500
BSR 101 .763 .658 .711 .605 .605 .711 .684 .684 .684
7260 .605 .605 .605 .658 .658 .658 .579 .632 .368
8252 .579 .684 .632 .737 .526 .632 .605 .605 .500
885100 .605 .605 .500 .605 .553 .658 .579 .395 .474
885166 .579 .684 .474 .684 .632 .737 .763 .395 .500
885168 .684 .684 .526 .632 .632 .737 .605 .395 .500
886348 .737 .737 .684 .684 .684 .684 .605 .447 .553
HARDIN .974 .553 .868 .500 .605 .605 .526 .553 .632
M84-916 .763 .500 .605 .447 .605 .605 .526 .632 .763
CHAPMAN .632 .842 .579 .790 .684 .790 .763 .500 .553
888080 .684 .790 .526 .790 .790 .684 .658 .605 .474
ELGIN .684 .790 .526 .737 .842 .842 .658 .553 .605
884098 .842 .632 .790 .579 .684 .579 .500 .658 .500
886339 .658 .763 .553 .711 .816 .763 .579 .500 .579
KENWOOD .629 .840 .476 .787 .892 .787 .603 .605 .550
A2234 .526 .947 .474 1.000 .790 .684 .711 .605 .447
ABS-293033 .684 .421 .632 .474 .579 .579 .395 .711 .500

887223 .632 .632 .526 .684 .632 .579 .447 .632 .500

64
Appendix C. continued

A83-271010 BSR 201 7299 J-231 PELLA 9271 9292 STURDY A2943
CORSOY .500 .632 .658 .790 .526 .579 .658 .763 .737
CWI‘URY .658 .737 .658 .684 .842 .579 .658 .763 .684
AP 1989 .447 .579 .711 .632 .579 .526 .553 .763 .632
HW8008 .605 .684 .658 .632 .790 .526 .605 .711 .632
82596 .658 .737 .605 .737 .684 .737 .816 .605 .632
A1937 .658 .737 .500 .737 .790 .737 .711 .605 .579
BURLISON .632 .605 .526 .553 .868 .500 .526 .632 .500
823-12 .790 .500 .632 .553 .395 .605 .658 .553 .368
SIBLEY .711 .368 .474 .711 .447 .447 .526 .579 .447
A83-271010 .974 .632 .526 .658 .605 .711 .684 .632 .447
BSR 201 .632 .947 .711 .632 .684 .842 .763 .711 .632
7299 .526 .711 .974 .658 .605 .553 .579 .684 .605
J-231 .658 .632 .658 1.000 .632 .579 .658 .763 .737
PELLA .605 .684 .605 .632 1.000 .526 .500 .711 .579
9271 .711 .842 .553 .579 .526 1.000 .921 .605 .526
9292 .684 .763 .579 .658 .500 .921 .974 .658 .579
STURDY .632 .711 .684 .763 .711 .605 .658 .974 .737
A2943 .447 .632 .605 .737 .579 .526 .579 .737 .947
A80-l47002 .763 .526 .421 .658 .711 .500 .474 .632 .605
A82—161035 .895 .605 .500 .684 .526 .684 .632 .579 .500
A84-185032 .632 .763 .684 .605 .447 .816 .737 .632 .605
ABS-291010 .763 .737 .711 .842 .579 .737 .816 .711 .632
1250 .553 .790 .763 .526 .474 .737 .658 .658 .579
CONRAD .632 .658 .526 .605 .763 .658 .711 .711 .579
C1664 .763 .737 .500 .632 .790 .632 .605 .711 .474
DSR-262 .553 .842 .658 .684 .842 .684 .605 .711 .684
884150 .684 .763 .632 .658 .658 .763 .842 .790 .658
884159 .737 .711 .553 .526 .526 .790 .711 .605 .632
884165 .658 .684 .632 .658 .500 .711 .737 .790 .658
HACK .632 .447 .605 .684 .632 .474 .553 .553 .526
LN82-296 .658 .737 .605 .632 .790 .632 .711 .711 .632
M82-946 .526 .421 .526 .711 .342 .553 .632 .526 .605
81884 .711 .579 .526 .711 .553 .553 .526 .579 .553
819-90 .500 .474 .605 .737 .790 .421 .500 .605 .579
ZANE .763 .737 .658 .790 .842 .579 .553 .868 .632
ABS-192034 .684 .711 .605 .763 .553 .711 .790 .579 .553
A85-292023 .500 .632 .500 .579 .421 .579 .658 .553 .684
A86-102004 .737 .711 .474 .658 .816 .711 .684 .632 .526
A86-103017 .632 .711 .684 .658 .658 .658 .737 .790 .605
A86-103027 .658 .632 .605 .632 .579 .737 .763 .553 .421
A86-202026 .632 .658 .684 .605 .605 .605 .684 .632 .447
A87-198005 .526 .658 .684 .763 .711 .605 .658 .658 .632
A87-296011 .605 .684 .553 .684 .632 .684 .711 .711 .579
A87-297015 .658 .737 .605 .632 .684 .737 .658 .553 .474
BSR 101 .737 .658 .632 .658 .711 .605 .579 .737 .500
7260 .632 .816 .790 .605 .605 .711 .632 .632 .553
8252 .553 .632 .711 .579 .737 .474 .395 .711 .579
885100 .526 .658 .447 .605 .658 .658 .605 .605 .684
885166 .553 .632 .526 .684 .790 .474 .447 .553 .632
885168 .605 .737 .500 .684 .737 .737 .711 .711 .684
886348 .500 .790 .711 .842 .684 .632 .658 .763 .737
HARDIN .474 .658 .684 .763 .500 .605 .658 .763 .737
M84-916 .711 .526 .474 .816 .447 .658 .737 .684 .658
CHAPMAN .711 .790 .605 .737 .895 .632 .605 .816 .579
888080 .632 .790 .684 .684 .632 .684 .763 .711 .632
ELGIN .658 .737 .658 .895 .737 .684 .763 .763 .632
884098 .553 .790 .816 .632 .474 .737 .763 .763 .684
886339 .579 .711 .632 .868 .658 .658 .711 .658 .711
KENWOOD .708 .792 .713 .840 .682 .740 .816 .711 .632
A2234 .605 .684 .658 .632 .790 .526 .605 .711 .632
ABS—293033 .605 .684 .658 .579 .368 .737 .658 .553 .526

887223 .605 1 .632 .658 .684 .526 .684 .737 .737 .684

6 5
Appendix C. continued

A80- A82- A84- A85-
147002 161035 185032 291010 1250 CONRAD C1664 DSR-262 884150
CORSOY .605 .526 .763 .737 .737 .605 .632 .579 .658
CENTURY .658 .579 .500 .737 .526 .816 .737 .790 .816
AP 1989 .553 .526 .711 .579 .684 .605 .579 .526 .553
HW8008 .605 .526 .447 .684 .579 .763 .684 .737 .763
82596 .553 .579 .605 .895 .579 .763 .684 .737 .711
A1937 .658 .579 .605 .790 .579 .763 .790 .842 .605
BURLISON .737 .553 .421 .605 .447 .790 .868 .711 .579
823-12 .579 .737 .632 .711 .658 .605 .605 .395 .579
SIBLEY .711 .737 .526 .605 .395 .579 .605 .395 .526
A83-271010 .763 .895 .632 .763 .553 .632 .763 .553 .684
BSR 201 .526 .605 .763 .737 .790 .658 .737 .842 .763
7299 .421 .500 .684 .711 .763 .526 .500 .658 .632
J-231 .658 .684 .605 .842 .526 .605 .632 .684 .658
PELLA .711 .526 .447 .579 .474 .763 .790 .842 .658
9271 .500 .684 .816 .737 .737 .658 .632 .684 .763
9292 .474 .632 .737 .816 .658 .711 .605 .605 .842
STURDY .632 .579 .632 .711 .658 .711 .711 .711 .790
A2943 .605 .500 .605 .632 .579 .579 .474 .684 .658
A80-147002 .974 .790 .526 .553 .447 .737 .763 .658 .579
A82-161035 .790 .947 .658 .684 .526 .579 .684 .526 .605
A84-185032 .526 .658 .921 .711 .868 .526 .553 .605 .632
A85-291010 .553 .684 .711 1.000 .684 .658 .684 .632 .711
1250 .447 .526 .868 .684 1.000 .500 .579 .632 .605
CONRAD .737 .579 .526 .658 .500 .974 .763 .763 .737
C1664 .763 .684 .553 .684 .579 .763 1.000 .737 .658
BSR-262 .658 .526 .605 .632 .632 .763 .737 1.000 .658
884150 .579 .605 .632 .711 .605 .737 .658 .658 .974
884159 .684 .711 .737 .632 .684 .605 .579 .684 .658
884165 .658 .632 .684 .658 .658 .684 .658 .658 .737
HACK .632 .553 .421 .684 .474 .605 .579 .579 .500
LN82-296 .658 .579 .553 .790 .579 .868 .790 .737 .763
M82-946 .526 .553 .632 .658 .605 .526 .447 .395 .526
81884 .816 .737 .579 .605 .605 .579 .711 .605 .579
819-90 .605 .526 .447 .579 .368 .658 .579 .632 .553
ZANE .763 .684 .605 .737 .632 .711 .842 .790 .711
A85-192034 .526 .658 .605 .868 .605 .632 .711 .605 .684
ABS-292023 .500 .526 .553 .632 .632 .500 .632 .474 .658
A86-102004 .684 .658 .579 .763 .553 .790 .816 .763 .632
A86-103017 .474 .553 .658 .711 .658 .632 .658 .605 .842
A86-103027 .553 .632 .553 .684 .474 .711 .684 .632 .605
A86-202026 .474 .553 .605 .763 .711 .684 .711 .553 .684
A87-198005 .632 .579 .526 .605 .447 .711 .605 .711 .737
A87-296011 .711 .684 .605 .632 .526 .868 .737 .684 .711
A87-297015 .553 .737 .658 .684 .579 .605 .684 .632 .553
BSR 101 .737 .658 .711 .711 .711 .737 .816 .658 .632
7260 .526 .605 .737 .763 .868 .579 .658 .711 .579
8252 .658 .526 .605 .526 .737 .658 .632 .790 .500
885100 .737 .579 .579 .447 .553 .658 .605 .763 .684
885166 .763 .579 .447 .526 .474 .605 .737 .790 .553
885168 .711 .579 .632 .579 .579 .711 .684 .790 .816
886348 .605 .579 .605 .684 .579 .711 .684 .842 .711
HARDIN .579 .526 .790 .711 .763 .605 .605 .605 .632
M84-916 .711 .737 .632 .711 .500 .579 .605 .500 .684
CHAPMAN .711 .632 .553 .684 .579 .763 .895 .842 .763
888080 .526 .553 .605 .790 .684 .658 .737 .684 .816
ELGIN .553 .579 .553 .842 .526 .711 .737 .790 .763
884098 .447 .526 .868 .790 .895 .605 .579 .632 .711
886339 .579 .632 .579 .763 .500 .658 .605 .763 .684
KENWOOD .497 .632 .608 .892 .582 .658 .682 .740 .813
A2234 .605 .526 .447 .684 .579 .763 .684 .737 .763
A85-293033 .447 .632 .842 .684 .895 .395 .526 .526 .500

887223 .553 .553 .632 .684 .684 .632 .474 .632 .763

66
Appendix C. continued

884159 884165 HACK 11482-296 M82-946 81884 819-90 ZANE ABS-192034

CORSOY .526 .711 .526 .632 .763 .658 .632 .684 .658
CENHURY .579 .553 .579 .947 .395 .500 .737 .790 .711
AP 1989 .474 .658 .474 .579 .605 .500 .684 .632 .500
HW8008 .526 .500 .632 .895 .447 .553 .684 .737 .711
82596 .632 .553 .684 .895 .553 .500 .684 .632 .868
A1937 .632 .553 .684 .790 .553 .605 .684 .737 .763
BURLISON .553 .579 .711 .816 .368 .579 .711 .711 .632
823-12 .605 .632 .658 .605 .684 .632 .500 .553 .684
SIBLEY .474 .605 .579 .500 .684 .658 .658 .605 .579
A83-271010 .737 .658 .632 .658 .526 .711 .500 .763 .684
BSR 201 .711 .684 .447 .737 .421 .579 .474 .737 .711
7299 .553 .632 .605 .605 .526 .526 .605 .658 .605
J-23l .526 .658 .684 .632 .711 .711 .737 .790 .763
PELLA .526 .500 .632 .790 .342 .553 .790 .842 .553
9271 .790 .711 .474 .632 .553 .553 .421 .579 .711
9292 .711 .737 .553 .711 .632 .526 .500 .553 .790
STURDY .605 .790 .553 .711 .526 .579 .605 .868 .579
A2943 .632 .658 .526 .632 .605 .553 .579 .632 .553
A80-147002 .684 .658 .632 .658 .526 .816 .605 .763 .526
A82-161035 .711 .632 .553 .579 .553 .737 .526 .684 .658
A84-185032 .737 .684 .421 .553 .632 .579 .447 .605 .605
ABS-291010 .632 .658 .684 .790 .658 .605 .579 .737 .868
1250 .684 .658 .474 .579 .605 .605 .368 .632 .605
CONRAD .605 .684 .605 .868 .526 .579 .658 .711 .632
C1664 .579 .658 .579 .790 .447 .711 .579 .842 .711
DSR-262 .684 .658 .579 .737 .395 .605 .632 .790 .605
884150 .658 .737 .500 .763 .526 .579 .553 .711 .684
884159 .947 .790 .605 .579 .447 .632 .368 .579 .553
884165 .790 .974 .632 .553 .579 .711 .395 .658 .579
HACK .605 .632 .947 .579 .605 .684 .632 .579 .658
LN82-296 .579 .553 .579 1.000 .447 .500 .684 .737 .763
M82-946 .447 .579 .605 .447 .974 .684 .553 .500 .632
81884 .632 .711 .684 .500 .684 .974 .447 .711 .632
819-90 .368 .395 .632 .684 .553 .447 1.000 .632 .605
ZANE .579 .658 .579 .737 .500 .711 .632 1.000 .605
ABS-192034 .553 .579 .658 .763 .632 .632 .605 .605 .921
ABS-292023 .526 .605 .526 .632 .658 .658 .421 .474 .763
A86-102004 .605 .474 .605 .868 .474 .526 .711 .763 .737
A86-103017 .526 .579 .447 .763 .526 .421 .711 .711 .684
A86-103027 .632 .763 .737 .579 .553 .605 .579 .526 .711
A86-202026 .421 .474 .553 .816 .632 .526 .658 .658 .790
A87-198005 .500 .684 .605 .605 .579 .684 .711 .658 .632
A87-296011 .526 .711 .474 .737 .605 .658 .632 .684 .658
A87-297015 .526 .447 .474 .684 .447 .500 .684 .632 .711
BSR 101 .553 .632 .526 .711 .632 .684 .605 .868 .579
7260 .658 .632 .605 .658 .526 .684 .395 .711 .684
8252 .579 .553 .579 .632 .500 .605 .632 .790 .447
885100 .711 .684 .553 .553 .526 .737 .553 .605 .526
885166 .579 .553 .684 .632 .447 .763 .684 .684 .632
885168 .684 .711 .526 .632 .553 .711 .579 .737 .579
886348 .526 .711 .526 .684 .553 .658 .684 .737 .711
HARDIN .553 .737 .500 .605 .737 .632 .605 .658 .632
M84-916 .684 .816 .684 .500 .737 .763 .553 .605 .684
CHAPMAN .526 .605 .526 .790 .447 .658 .684 .947 .658
888080 .605 .684 .605 .790 .526 .632 .526 .684 .816
EIGIN .526 .658 .684 .737 .605 .605 .737 .790 .816
884098 .684 .763 .474 .684 .605 .500 .474 .632 .658
886339 .553 .579 .605 .711 .579 .579 .763 .658 .790
KENWOOD .582 .608 .629 .787 .550 .550 .682 .734 .866
A2234 .526 .500 .632 .895 .447 .553 .684 .737 .711
ABS-293033 .684 .605 .526 .474 .658 .605 .421 .526 .658

887223 .711 .763 .684 .579 .658 .658 .474 .632 .605

6 7
Appendix C. Continued

A85- A86- A86- A86- A86- A87- A87- A87-
292023 102004 103017 103027 202026 198005 296011 297015 BSR 101

CORSOY .684 .553 .711 .526 .658 .658 .684 .526 .763
CENTURY .579 .816 .816 .526 .763 .658 .684 .632 .658
AP 1989 .526 .500 .658 .579 .605 .658 .737 .684 .711
HW8008 .632 .763 .763 .474 .816 .605 .632 .579 .605
82596 .632 .868 .711 .684 .763 .605 .632 .684 .605
A1937 .526 .921 .605 .684 .658 .605 .632 .684 .711
BURLISON .553 .790 .579 .711 .632 .632 .658 .658 .684
823-12 .553 .579 .632 .658 .790 .447 .605 .553 .684
SIBLEY .447 .526 .579 .605 .579 .579 .658 .500 .684
A83-271010 .500 .737 .632 .658 .632 .526 .605 .658 .737
BSR 201 .632 .711 .711 .632 .658 .658 .684 .737 .658
7299 .500 .474 .684 .605 .684 .684 .553 .605 .632
J-231 .579 .658 .658 .632 .605 .763 .684 .632 .658
PELLA .421 .816 .658 .579 .605 .711 .632 .684 .711
9271 .579 .711 .658 .737 .605 .605 .684 .737 .605
9292 .658 .684 .737 .763 .684 .658 .711 .658 .579
STURDY .553 .632 .790 .553 .632 .658 .711 .553 .737
A2943 .684 .526 .605 .421 .447 .632 .579 .474 .500
A80-147002 .500 .684 .474 .553 .474 .632 .711 .553 .737
A82-161035 .526 .658 .553 .632 .553 .579 .684 .737 .658
A84-185032 .553 .579 .658 .553 .605 .526 .605 .658 .711
ABS-291010 .632 .763 .711 .684 .763 .605 .632 .684 .711
1250 .632 .553 .658 .474 .711 .447 .526 .579 .711
CONRAD .500 .790 .632 .711 .684 .711 .868 .605 .737
C1664 .632 .816 .658 .684 .711 .605 .737 .684 .816
DSR-262 .474 .763 .605 .632 .553 .711 .684 .632 .658
884150 .658 .632 .842 .605 .684 .737 .711 .553 .632
884159 .526 .605 .526 .632 .421 .500 .526 .526 .553
884165 .605 .474 .579 .763 .474 .684 .711 .447 .632
HACK .526 .605 .447 .737 .553 .605 .474 .474 .526
LN82-296 .632 .868 .763 .579 .816 .605 .737 .684 .711
M82-946 .658 .474 .526 .553 .632 .579 .605 .447 .632
81884 .658 .526 .421 .605 .526 .684 .658 .500 .684
819-90 .421 .711 .711 .579 .658 .711 .632 .684 .605
ZANE .474 .763 .711 .526 .658 .658 .684 .632 .868
ABS-192034 .763 .737 .684 .711 .790 .632 .658 .711 .579
ABS-292023 1.000 .500 .605 .526 .658 .553 .579 .526 .447
A86-102004 .500 .974 .684 .605 .737 .526 .658 .763 .737
A86-103017 .605 .684 .974 .500 .816 .579 .605 .605 .684
A86-103027 .526 .605 .500 1.000 .553 .763 .737 .684 .553
A86-202026 .658 .737 .816 .553 .974 .526 .658 .658 .737
A87-198005 .553 .526 .579 .763 .526 .974 .816 .658 .579
A87-296011 .579 .658 .605 .737 .658 .816 1.000 .737 .711
A87-297015 .526 .763 .605 .684 .658 .658 .737 1.000 .605
BSR 101 .447 .737 .684 .553 .737 .579 .711 .605 .974
7260 .605 .632 .526 .605 .684 .579 .605 .711 .684
8252 .368 .658 .605 .421 .658 .500 .579 .526 .763
885100 .605 .579 .526 .553 .421 .711 .658 .500 .526
885166 .632 .658 .500 .579 .500 .711 .579 .579 .553
885168 .579 .658 .658 .579 .500 .763 .684 .526 .658
886348 .632 .605 .658 .684 .605 .868 .842 .684 .605
HARDIN .658 .526 .684 .553 .632 .658 .711 .553 .737
M84-916 .658 .526 .579 .711 .474 .684 .658 .500 .579
CHAPMAN .526 .816 .763 .579 .711 .711 .737 .684 .816
888080 .790 .658 .763 .632 .763 .658 .632 .579 .605
ELGIN .579 .763 .763 .737 .711 .763 .684 .632 .658
884098 .632 .553 .763 .579 .711 .553 .632 .579 .711
886339 .605 .711 .684 .658 .632 .763 .711 .711 .526
KENWOOD .629 .761 .813 .687 .761 .711 .634 .687 .603
A2234 .632 .763 .763 .474 .816 .605 .632 .579 .605
A85-293033 .632 .553 .605 .526 .658 .395 .474 .632 .605

887223 .579 .526 .632 .579 .579 .632 .579 .421 .553

68
Appendix C. continued

7260 8252 885100 885166 885168 886348 HARDIN M84-916 CHAPMAN
CORSOY .605 .579 .605 .579 .684 .737 .974 .763 .632
CBPI‘URY .605 .684 .605 .684 .684 .737 .553 .500 .842
AP 1989 .605 .632 .500 .474 .526 .684 .868 .605 .579
HW8008 .658 .737 .605 .684 .632 .684 .500 .447 .790
82596 .658 .526 .553 .632 .632 .684 .605 .605 .684
A1937 .658 .632 .658 .737 .737 .684 .605 .605 .790
BURLISON .579 .605 .579 .763 .605 .605 .526 .526 .763
823-12 .632 .605 .395 .395 .395 .447 .553 .632 .500
SIBLEY .368 .500 .474 .500 .500 .553 .632 .763 .553
A83-271010 .632 .553 .526 .553 .605 .500 .474 .711 .711
BSR 201 .816 .632 .658 .632 .737 .790 .658 .526 .790
7299 .790 .711 .447 .526 .500 .711 .684 .474 .605
J-231 .605 .579 .605 .684 .684 .842 .763 .816 .737
PELLA .605 .737 .658 .790 .737 .684 .500 .447 .895
9271 .711 .474 .658 .474 .737 .632 .605 .658 .632
9292 .632 .395 .605 .447 .711 .658 .658 .737 .605
STURDY .632 .711 .605 .553 .711 .763 .763 .684 .816
A2943 .553 .579 .684 .632 .684 .737 .737 .658 .579
A80-147002 .526 .658 .737 .763 .711 .605 .579 .711 .711
A82-161035 .605 .526 .579 .579 .579 .579 .526 .737 .632
A84-185032 .737 .605 .579 .447 .632 .605 .790 .632 .553
ABS-291010 .763 .526 .447 .526 .579 .684 .711 .711 .684
1250 .868 .737 .553 .474 .579 .579 .763 .500 .579
CONRAD .579 .658 .658 .605 .711 .711 .605 .579 .763
C1664 .658 .632 .605 .737 .684 .684 .605 .605 .895
BSR-262 .711 .790 .763 .790 .790 .842 .605 .500 .842
884150 .579 .500 .684 .553 .816 .711 .632 .684 .763
884159 .658 .579 .711 .579 .684 .526 .553 .684 .526
884165 .632 .553 .684 .553 .711 .711 .737 .816 .605
HACK .605 .579 .553 .684 .526 .526 .500 .684 .526
LN82—296 .658 .632 .553 .632 .632 .684 .605 .500 .790
M82-946 .526 .500 .526 .447 .553 .553 .737 .737 .447
81884 .684 .605 .737 .763 .711 .658 .632 .763 .658
819-90 .395 .632 .553 .684 .579 .684 .605 .553 .684
ZANE .711 .790 .605 .684 .737 .737 .658 .605 .947
A85-192034 .684 .447 .526 .632 .579 .711 .632 .684 .658
ABS-292023 .605 .368 .605 .632 .579 .632 .658 .658 .526
A86-102004 .632 .658 .579 .658 .658 .605 .526 .526 .816
A86-103017 .526 .605 .526 .500 .658 .658 .684 .579 .763
A86-103027 .605 .421 .553 .579 .579 .684 .553 .711 .579
A86-202026 .684 .658 .421 .500 .500 .605 .632 .474 .711
A87-198005 .579 .500 .711 .711 .763 .868 .658 .684 .711
A87-296011 .605 .579 .658 .579 .684 .842 .711 .658 .737
A87-297015 .711 .526 .500 .579 .526 .684 .553 .500 .684
BSR 101 .684 .763 .526 .553 .658 .605 .737 .579 .816
7260 .974 .711 .526 .579 .553 .658 .632 .474 .658
8252 .711 1.000 .605 .632 .579 .632 .605 .395 .737
885100 .526 .605 .921 .790 .868 .711 .605 .684 .658
885166 .579 .632 .790 .947 .737 .737 .553 .605 .737
885168 .553 .579 .868 .737 .947 .737 .658 .711 .790
886348 .658 .632 .711 .737 .737 1.000 .763 .658 ”790
HARDIN .632 .605 .605 .553 .658 .763 .974 .737 .605
M84-916 .474 .395 .684 .605 1711 .658 .737 .974 .553
CHAPMAN .658 .737 .658 .737 .790 .790 .605 .553 1.000
888080 .711 .526 .605 .658 .684 .737 .658 .632 .737
BEGIN .605 .579 .605 .684 .737 .842 .658 .711 .842
884098 .763 .632 .500 .421 .579 .684 .868 .605 .579
886339 .579 .553 .658 .711 .684 .868 .658 .684 .711
KENWOOD .661 .529 .553 .629 .682 .792 .605 .655 .787
A2234 .658 .737 .605 .684 .632 .684 .500 .447 .790
A85-293033 .763 .632 .526 .474 .500 .526 .711 .605 .474

887223 .658 .632 .684 .526 .711 .632 .632 .711 .579

69
Appendix C. continued

888080 ELGIN 884098 886339 KENWOOD A2234 ABS-293033 887223

CORSOY .684 .684 .842 .658 .629 .526 .684 .632
CENTURY .790 .790 .632 .763 .840 .947 .421 .632
AP 1989 .526 .526 .790 .553 .476 .474 .632 .526
HW8008 .790 .737 .579 .711 .787 1.000 .474 .684
82596 .790 .842 .684 .816 .892 .790 .579 .632
A1937 .684 .842 .579 .763 .787 .684 .579 .579
BURLISON .658 .658 .500 .579 .603 .711 .395 .447
823-12 .605 .553 .658 .500 .605 .605 .711 .632
SIBLEY .474 .605 .500 .579 .550 .447 .500 .500
A83-271010 .632 .658 .553 .579 .708 .605 .605 .605
BSR 201 .790 .737 .790 .711 .792 .684 .684 .632
7299 .684 .658 .816 .632 .713 .658 .658 .658
J-231 .684 .895 .632 .868 .840 .632 .579 .684
PELLA .632 .737 .474 .658 .682 .790 .368 .526
9271 .684 .684 .737 .658 .740 .526 .737 .684
9292 .763 .763 .763 .711 .816 .605 .658 .737
STURDY .711 .763 .763 .658 .711 .711 .553 .737
A2943 .632 .632 .684 .711 .632 .632 .526 .684
A80-147002 .526 .553 .447 .579 .497 .605 .447 .553
A82-161035 .553 .579 .526 .632 .632 .526 .632 .553
A84-185032 .605 .553 .868 .579 .608 .447 .842 .632
ABS-291010 .790 .842 .790 .763 .892 .684 .684 .684
1250 .684 .526 .895 .500 .582 .579 .895 .684
CONRAD .658 .711 .605 .658 .658 .763 .395 .632
C1664 .737 .737 .579 .605 .682 .684 .526 .474
BSR-262 .684 .790 .632 .763 .740 .737 .526 .632
884150 .816 .763 .711 .684 .813 .763 .500 .763
884159 .605 .526 .684 .553 .582 .526 .684 .711
884165 .684 .658 .763 .579 .608 .500 .605 .763
HACK .605 .684 .474 .605 .629 .632 .526 .684
LN82-296 .790 .737 .684 .711 .787 .895 .474 .579
M82-946 .526 .605 .605 .579 .550 .447 .658 .658
81884 .632 .605 .500 .579 .550 .553 .605 .658
819-90 .526 .737 .474 .763 .682 .684 .421 .474
ZANE .684 .790 .632 .658 .734 .737 .526 .632
A85-192034 .816 .816 .658 .790 .866 .711 .658 .605
A85-292023 .790 .579 .632 .605 .629 .632 .632 .579
A86-102004 .658 .763 .553 .711 .761 .763 .553 .526
A86-103017 .763 .763 .763 .684 .813 .763 .605 .632
A86-103027 .632 .737 .579 .658 .687 .474 .526 .579
A86-202026 .763 .711 .711 .632 .761 .816 .658 .579
A87-198005 .658 .763 .553 .763 .711 .605 .395 .632
A87-296011 .632 .684 .632 .711 .634 .632 .474 .579
A87-297015 .579 .632 .579 .711 .687 .579 .632 .421
BSR 101 .605 .658 .711 .526 .603 .605 .605 .553
7260 .711 .605 .763 .579 .661 .658 .763 .658
8252 .526 .579 .632 .553 .529 .737 .632 .632
885100 .605 .605 .500 .658 .553 .605 .526 .684
885166 .658 .684 .421 .711 .629 .684 .474 .526
885168 .684 .737 .579 .684 .682 .632 .500 .711
886348 .737 .842 .684 .868 .792 .684 .526 .632
HARDIN .658 .658 .868 .658 .605 .500 .711 .632
M84-916 .632 .711 .605 .684 .655 .447 .605 .711
CHAPMAN .737 .842 .579 .711 .787 .790 .474 .579
888080 .895 .790 .737 .711 .840 .790 .605 .684
EIGIN .790 1.000 .632 .868 .945 .737 .526 .684
884098 .737 .632 1.000 .605 .687 .579 .790 .684
886339 .711 .868 .605 .921 .868 .711 .553 .632
KENWOOD .840 .945 .687 .868 .995 .787 .582 .684
A2234 .790 .737 .579 .711 .787 1.000 .474 .684
ABS—293033 .605 .526 .790 .553 .582 .474 .947 .632

887223 .684 .684 .684 .632 .684 .684 .632 .895

Application of Genetic Distances to a Soybean Breeding
Program

Abstract

The genetic distance was estimated between elite soybean lines
using the coefficient of parentage (CP). a similarity index (SI) based on
marker loci. and a distance value (PCD) derived from a principal
component analysis of CP and SI data. Genetically similar and distant
parents were crossed based on the SI values and approximately 55
82:3 or F4:5 families were derived from each cross and then field
tested for seed yield. plant height. and date of maturity. The
generalized variance. and the genetic and progeny variance for
individual traits were estimated for each population. In general the
variance parameters of a population increased as the genetic distance
between the parents of the population increased. regardless of the
method of estimating the distance. There was a stronger association
of the measures with the generalized variances than with the genetic
variances of the individual traits of the populations. No measure was
significantly associated with the yield genetic variance of a population
while all seemed predictive of the genetic variances for the maturity
and height traits. The utility of the measures appeared to be limited
to identifying which populations would have a higher likelihood of
having an above average genetic variance for the individual traits. The
PCD appeared to be the best predictor of the variance parameters of
the populations.

Introduction

The objective of breeding soybeans (Glycine max L. Merr.) for a
metric trait such as yield is to develop progeny from a segregating
population that are better than the parents of the population. The
mean of the progeny derived from selected individuals (u') can be
expressed in terms derived from the gain from mass selection
equation (Falconer. 1981)

u'=(us-u)h2+u

where u is the mean of the parental population. (us - u) is the selection
differential. us is the mean of the selected individuals. and h2 is the
ratio of additive genetic variance to phenotypic variance. Assuming
equal selection differentials then u' will increase as u and h2 increase.
A high u can be assured by crossing elite parents that have an
acceptable level of expression for the trait to be improved and b2 can
be improved in a population by crossing parents that each posses
different alleles at the loci that control the trait. Elite parents can be
identified through testing and their use as parents is a common
practice in cultivar development programs. A quantification of the
actual amount of allelic differences between two parents. defined as
their genetic distance. can be estimated but is rarely used in parent
selection as breeders usually cross lines that they instinctively deem as
distant.

There are various methods of estimating the genetic distance
between two individuals and each has it's own assumptions and biases
that could influence it's ability to accurately estimate the true genetic
distance. Genetic distance can be estimated from pedigree data by the
coefficient of parentage (CP) (Kempthorne. 1969) which is the
probability of sampling an allele fi'om one individual that is identical by
descent to an allele sampled from the same locus in the other
individual. The value (1 - CP) estimates the genetic distance and the
ability of this value to predict the true genetic distance between
individuals is entirely dependent on the validity of the assumptions
that are made during it's calculation. There are extensive pedigree

7 O

71

records for adapted soybeans that would allow this approach to be
used with an elite population. Another approach is to use highly
heritable phenotypes, such as isozymes. restriction fragment length
polymorphisms (RFLPs). and certain morphological traits. as
qualitative markers to assay for differences between parents at the
DNA level. The loci the control such polymorphisms are termed
marker loci (ML). The frequency of each polymorphism at each ML in
each parent is determined and a comparison equation is employed to
estimate the genetic distance between parents. The accuracy of the
estimate will depend on obtaining a representative sample of
potentially polymorphic loci and on the assumption that alleles that
produce identical phenotypes are themselves identical. Suitable
polymorphisms occur with a moderate frequency in adapted soybeans
(Doong and Kiang. 1987; Gorman et aL 1983; Kiem et al. 1989)
making this method a practical option. A third approach infers
genetic distance between parents from a statistical distance based on
quantitative trait differences (Goodman. 1972) and can assay for allelic
differences at many loci when truly polygenic traits are measured.
Some limitations of this approach are that each quantitative phenotype
assays an unknown number of genetic differences that may not be
proportional to the observed phenotypic differences and the effect of
the environment on the traits requires that each trait and genotype be
assayed in multiple locations and years. There is a restricted range of
phenotypes among the elite soybean lines that might limit the utility of
this approach within this population

There has been extensive use of genetic distances to estimate
the relationship among individuals and yet there has been relatively
little research investigating whether estimates of parental genetic
distance relate to the parameters of a hybrid or segregating population
that would be expected to be influenced by this distance and that
concern a breeder. Such parameters include the amount of heterosis.
the number of transgressive segregants. or the amount of genetic
variance (Cowen and Frey, 1987). A low. though significant, positive
correlation between parental genetic distances estimated with CP and
hybrid mid-parent heterosis for grain yield was reported in rapeseed
(Lefort-Buson et al., 1987) while the association was not significant in

72

oats (Cowen and Frey. 1987) or wheat though Cox and Murphy (1990)
noted the value of using the CF to select diverse wheat parents after
the parental candidates had already been selected for their
performance and phenotypic divergence. Cowen and Frey (1987)
found a significant association between the CP of the parents of a
segregating oat population and the number of transgressive segregants
for plant height but not for the traits bundle weight. grain yield. straw
yield, heading date. and harvest index.. They also reported a
significant association between the CP of the parents and the
generalized genetic variance of the segregating population.

There have been several attempts. with mixed results. to
correlate a ML derived parental genetic distance with the
performance of hybrids between inbred maize lines. which is generally
dependent on the extent of heterosis. Price et al. (1986) and Lamkey
et al. (1987) found a significant but low positive correlation between
hybrid grain yield and inbred parental genetic distances based on
isozymes indicating that the diversity estimate had little value in
predicting hybrid performance. Other researchers using isozymes
(Hunter and Kannenberg, 1971: Hadjinov et al. 1982) and RFLPs
(Godshalk et al.. 1990) have found insignificant correlations between
hybrid grain yield of maize and parental genetic distances. Frei et al.
(1986) found the ML genetic distance between parents to be
predictive of hybrid performance only when there was a known.
pedigree relationship between the parents: a situation that increases
inbreeding and the probability of linkage disequilibrium between ML
and other genes. Lee et al (1989). using RFLPs to estimate parental
diversity within and across chromosomes. found a significant
correlation between parental genetic distance and hybrid grain yield of
maize and that the diversity of some chromosomes was more
predictive of yield than other chromosomes suggesting that this
approach would be more successful when only the diversity of relevant
chromosome segments is used. No research has been performed
relating ML derived genetic distances to the amount of genetic
variance or the number of transgressive segregants in segregating
populations.

73

The purpose of the reported research was to investigate if a CP
or a ML based estimate of genetic distance between parents could be
used to to predict which elite soybean lines. when crossed. would
produce the greatest amount of useful genetic variability in the
resulting segregating population.

Materials and Methods

Sixty-two public breeding lines and public and private cultivars
were selected as elite parents on the basis of their high yield and
adapted maturity in performance trials conducted over several years by
Michigan State University. The genotype of each elite line was
determined for the following five morphological and thirteen isozyme
loci: the W, T, L, I. and R loci that control the color of the flower,
pubescence, pod. hylum. and hylum respectively. diaphorase (Dtaj
locus; EC 1.6.4.3). endopeptidase (Enp locus). isocitrate
dehydrogenase (Idhl, Idhz loci; EC 1.1.1.42), mannose phosphate
isomerase (Mpi locus: EC 5.10.11). phosphoglucosemutase (Pgmj
locus: EC 5.3.1.9). and superoxide dimutase (Sod locus: EC 1.15.1.1)
acid phosphatase (Acp locus: EC 3.1.3.2). aconitase (Acoz. A004 loci;
EC 4.2.1.3). fluorescent esterase (Fle locus), 6-phosphogluconate
dehydrogenase (Pgd; locus: EC 1.1.1.44). and phosphoglucose
isomerase (Pgt locus; EC). The isozyme methodology and results
reported in Sneller (1991). The level of peroxidase activity was also
ascertained according to the method of Buttery and Buzzel (1968).
The data from these loci were used to calculate a similarity index (SI)
that is the probability of sampling an allele from one individual that is
identical by state to an allele sampled from the same locus in the other
individual. averaged over all loci (Sokal and Sneath, 1963). The S1
value can range from unity for two homozygous, homogeneous lines
that contain the same alleles to zero for lines that have no alleles in
common. The SI was calculated between all pairs of the 62 elite lines
using a program written in SAS Interactive Matrix Language by the
author and was used to select similar and distant parents which were
then crossed to produce the segregating populations that were
analyzed in this study (Table 1).

74

The C? was calculated between the 62 elite lines according to
the formula of Kempthorne (1969). Derivations of this formula
(Sneller. 1991) were used to calculated the CP between any genotype
and "Elgin” and "Kenwood" which were derived from the AP6 soybean
population (Fehr and Ortiz. 1978) that was created by an essentially
random mating of 40 genotypes. The following assumptions were
made in all CP calculations: (1) all the ancestral parents were
completely unrelated, including lines selected from other ancestral
parents, (ii) all ancestral parents were completely homogeneous and
inbred. and (iii) that each parent of a biparental cross contributed
equally to all progeny derived from the cross. A genotype derived by
five or more backcrosses was considered to be genetically equivalent
to the recurrent parent. The CP between the parents that were
crossed to generate the populations used in this study are listed in
Table 1. All CPs were calculated using programs written by the author
in SAS Interactive Matrix Language and a complete summary of the
CPs among the elite lines can be found in Sneller (1991)

A distance estimate was calculated that encompassed both the
CP and marker loci data. The C? was calculated between each of the
62 elite lines and each of the n ancestral parents that contributed
genes to the elite population resulting in a data set where each elite
line was described by n variables. This data set was combined with the
marker loci data set and a principal component analysis was
performed on the correlation matrix. The principal component
distance (PCD) was calculated between all elite lines according to the
formula

PCDg = [2 {(Yu. - ijfl / 3...} 11/2

where Y“. and Y3}. are the scores of the kth principal component for
the ith and jth elite lines and M is the eigenvalue of the kth principal
component (Goodman. 1972). Only principal components where M:
was greater than one were used in the calculation. The PCD between
the parents used to generate the populations analyzed in this study are
listed in Table 1. The PCDs were calculated using a program written

75

Table 1. Summary of the parentage. coefficient of parentage (CP),
similarity index (SI). and the principal component distance (PCD)
between the parents for all crosses used to generate the segregating
populations along with the number of families in each population and
the years tested.

 

 

Number of Year(s)
Cross Parent 1 Parent 2 CP SI PCD families tested

1 Hack Pella .48 .63 2.40 54 1990, 1989
2 Hardin A80-147002 .29 .58 3.01 69 1990, 1989
3 ABC-147002 A82-161035 .33 .90 1.88 60 1990, 1989
4 A83-271010 32596 .23 .66 3.02 58 1990. 1989
5 Burlison Century .59 .76 2.77 50 1990, 1989
6 C1664 Century .56 .74 2.19 66 1990, 1989
7 HW8008 Century .72 .95 2.23 55 1990, 1989
8 HW8008 ‘ LN82-296 .55 .90 1.70 54 1990, 1989
9 A2934 32596 .17 .63 5.02 58 1990. 1989
10 AGO-147002 E84098 .25 .45 3.64 47 1990. 1989
11 32596 A82-161035 .26 .58 3.34 63 1990, 1989
12 A1937 885100 .29 .66 1.64 51 1990

13 A83-271010 s18-84 .56 .71 2.79 54 1990

14 885168 A84-185032 .26 .63 1.97 55 1990

15 C1664 7260 .27 .66 2.38 54 1990

16 Hack 884150 .31 .50 3.79 53 1990

17 s23-12 Pella .10 .40 4.12 44 1990

18 Pella 885166 .43 .79 1.66 54 1990

19 323-12 1250 .17 .66 4.99 54 1990

20 1250 885166 .29 .47 3.82 52 1990

21 s23-12 885166 .16 .40 4.32 53 1990

22 Burlison M82-946 .19 .37 4.57 52 1990

23 E84165 Century .12 .55 3.14 60 1990

24 A2934 A83-271010 .18 .45 4.98 43 1990

 

76

in SAS Interactive Matrix Language by the author while the principal
component analysis was performed with standard SAS sofiware.

The progeny from eleven difi‘erent crosses were field tested in
both 1989 and 1990 (Table 1). The crosses were madeduring the

summer of 1987. the hybrids were grown in a greenhouse during the
winter of 1987-88 and the F2 populations were grown in the field

during the summer of 1988. An average of 59 82:3 families were
generated from each cross (Table 1). Each family of each cross was
field tested for seed yield. plant height. and date of maturity in single
row. 1.2 M long. hill plots. planted in 76.2 cm rows in a randomized
complete block design experiment with two replications. The seeding
rate was 27 seeds per meter of row. The experiment was planted and
harvested in Lenawee county. Michigan. in 1989 and in Ingham
county. Michigan. in 1990. An additional 13 crosses, each
represented by an average of 52 F4:5 families. plus the parents of all
crosses were similarly tested in 1990 only (Table l) at the Ingham
county location. These crosses were made in 1987. the hybrids grown
during the summer of 1988. and then advanced through two
generations of single seed descent during the winter of 1988-89 with
the F45 families being harvested from the 1989 summer planting of F4
seed. These crosses were added to the study to provide a broader
range of SI and CP values for the 1990 data analysis.

The variance among the progeny families of a cross for a
particular trait was estimated from a separate ANOVA of that trait for
each cross from the 1989. 1990. and the combined 1989-90 data sets.

The progeny variances (Ozprogeny) from the single years and from the
analysis combined over both years were interpreted as follows:

Single year 62131-093,” = 629 + mgxe
N0 year mprogeny = 629

where (1'29 and Ozgxe are the genetic and genotype by environment

variances of the population respectively, The generalized variance of
each progeny population was estimated by taking the determinate of

77

the variance-covariance matrix of the trait means from each family
member for the 1989, 1990 and 1989-1990 data sets. The natural log
of the generalized variance (GV) was used as an estimate of the overall
variability of each population (Goodman. 1968: Sokal. 1965). All
statistics were performed with SAS software.

Results

The generalized variances and progeny variances for the traits
seed yield. maturity. and height of the populations estimated in 1989.
1990 and 1989-90 are presented in Table 2. The estimate of the
progeny variance was negative in some cases and these values were
assumed to be zero in any further analyses. The CP. SI. and PCD
estimates of the genetic distance between the parents of the crosses
were significantly correlated to one another (Table 3) and each was
significantly correlated to the GVs of the populations derived from
these crosses regardless of the year in which the GVs were estimated.
The CP and SI measure the genetic similarity between two parents
and the significant negative correlations of these values with the (We
means that the overall variability of a population decreased as the
estimated genetic similarity of it's parents increased. The PCD
estimates parental genetic distance. not similarity. and this measure
was positively correlated with the CV values of the segregating
populations. Only linear effects were significant in the regressions of
the GVs of the populations on CP, S1 or PCD values between the
parents (Table 4). Each measure of parental genetic distance
accounted for a significant amount of the variation among the GVs of
the populations and all were approximately equal in their ability to
predict the CV of a population. The regressions of 1990 GVs on the
CP. 1989 GVs on the SI, and the 1989-90 GVs on the PCDs are shown
in Figures 1-3 respectively.

The CV of a population encompasses the genetic variation for all
traits. the genetic covariance between all traits. all non-genetic
sources of variation and covariation as well as the interactions between
these effects. The 1989 estimates of the GVs were significantly
correlated (r = 0.93) with the 1990 estimates suggesting that the CV

Table 2. Summary of the generalized variances (GV) and the maturity, height. and seed
yield progeny variances (PV) estimated in 1989. 1990 and 1989-90 for each

 

 

 

 

 

‘popmdatknr

GY M 11mm: pv __1;Lald W
Cross 1989 1990 1989-90 1989 1990 1989-909 1989 1990 1989-90 1989 1990 1989-90
1 12.1 10.7 9.9 15.4 1.6 3.6 2.0 0.5 2.2 2409.4 1684.1 0.0
2 13.7 13.0 12.5 36.8 7.2 16.8 16.1 11.0 3.4 7535.1 44.4 1659.1
3 11.0 11.0 9.8 4.4 1.5 1.9 1.0 2.8 1.3 1086.6 766.4 354.6
4 12.3 10.9 10.6 10.0 1.0 2.0 6.0 4.4 3.0 2280.6 1360.9 410.6
5 11.1 10.4 9.7 6.0 1.5 2.8 2.2 1.5 1.3 633.1 235.0 1096.2
6 10.6 10.3 9.2 4.7 1.6 2.1 0.6 0.1 0.0 1035.1 0.0 224.9
7 11.1 10.6 9.6 4.9 1.1 1.0 2.5 1.2 1.4 453.6 1333.8 306.5
8 11.1 10.2 9.3 5.1 1.2 2.2 0.7 0.0 1.7 495.3 0.0 26.1
9 13.4 12.4 12.0 33.0 9.2 15.6 6.9 8.3 9.0 6930.6 595.3 1214.6
10 13.7 12.9 12.0 23.4 10.1 8.8 10.0 5.3 5.2 5383.7 1307.1 12.4
11 12.1 11.3 10.7 8.1 1.6 2.3 5.3 5.5 2.9 516.1 806.3 574.9
12 10.8 . 0.0 7.1 . . 2113.0
13 10.8 0.3 3.5 . . 1122.5
14 10.3 0.2 5.6 . . 1201.3
15 11.1 1.3 3.2 1298.6 .
16 11.4 2.7 5.4 1105.5 .
17 11.8 3.5 6.5 486.9 .
18 11.3 2.5 . 1.0 . . 0.0
19 13.5 8.4 7.9 . . 2314.5
20 13.5 13.3 13.5 . 66.3
21 . 11.7 3.7 6.4 . 42.8
22 . 11.5 2.8 6.1 . 978.0
23 . 12.3 3.0 8.8 . . 1696.9
24 13.4 15.4 8.9 . . 2053.5

 

 

 

 

 

8‘ The 1989-90 progeny variance estimates the genetic variation within a population for
aspautumuknrtnan;

 

79

Table 3. Spearman's rank correlations of the generalized variances
(GV) and the maturity. height. and seed yield progeny variances (PV)
of the segregating populations with the coefficient of parentage (CP).
similarity index (81). and the principal component distance (PCD)
values of the parents of the populations. summarized by year(s).

 

 

 

 

 

 

Wre—
Year(s) CP SI PCD
cv 1989 -.72** -.6S*** .31...
CV 1990 .57eee -.79eee _73eee
cv 1989-1990 -.79*** -.81*** .78***
Maturity pv 1989 -.65** .80*** .77***
Maturity PV 1990 -.44 .78*** .56*
Maturity PV 1989-19909 .45 .77*** .60**
Height pv 1989 .65H .68" .84***
Height pv 1990 .73*** -.67*** .80***
Height pv 1989-1990 .77*** -.73** .75***
Yield PV 1989 -.67** .69“ .55*
Yield pv 1990 .24 .17 .33
Yield pv 1989-1990 .28 .13 .41

 

a The 1989-90 progeny variance estimates the genetic variation within
a population for a particular trait.

‘. “. “‘. denote significance at the a = 0.10. 0.05. and 0.01 levels
respectively. The 1990 correlations were estimated with 22 (if while
the 1989 and 1989-1990 correlations were estimated with 9 df.

80

Table 4. Summary of the R2 values and their significance from the
regressions of the GV and maturity, height and seed yield progeny
variances (PV) of the segregating populations on the coefficient of
parentage (CP). similarity (SI) and principal component distance
(PCD) values of the parents of the populations, summarized by year(s).

 

 

 

 

 

.__Lns&dnse_mensnze__.
Year (3) CP SI PCD
CV 1989 .56*** .66*** .58***
CV 1990 .37*** .32*** .53***
CV 1989-90 .59*** .57*** .61***
Maturity.PV 1989 .37** .45** .so**
Maturity.pv 1990 .17** .23** .47***
Maturity pv 1989-9oa .32* .31* .48**
Height pv 1989 .36* .47** .31*
Height pv 1990 .so*** .42*** .38***
Height pv 1989-90 .34* .29* .37**
Yield pv 1989 .41** .42** .49**
Yield FY 1990 .06 .01 .03
Yield FY 1989-90 .09 .05 .24

 

a The 1989-90 progeny variance estimates the genetic variance within
a population for a particular trait.

‘. “. “‘. denote significance of the linear regression model at the a
= 0.10, 0.05, and 0.01 levels respectively. The error of the 1990
regressions had 22 df while the 1989 and 1989-1990 regression
errors had 9 df.

81

 

14 l l I l I l r

13- - -

12

1990 GV

11

 

 

 

10
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

 

C?

Figure 1. Results of the linear regression of the 1990 estimates of the
generalized variances (GV) of the segregating populations on the
coefficient of parentage (CP) of the parents of the populations.

82

 

 

 

 

14
13 -
>
(D
a 12 -
Q
0‘
H
11 -
10 1 l 1 l 1 I

 

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
SI
Figure 2. Results of the linear regression of the 1989 estimates of the

generalized variances (GV) of the segregating populations on the
similarity index (SI) of the parents of the populations.

83

 

13

12-

1989-90 GV
H
H
I

10'-

 

 

 

 

PCD

Figure 3. Results of the linear regression of the 1989-90 estimates of
the generalized variances (GV) of the segregating populations on the
principal component distance (PCD) between the parents of the
populations.

84

was a repeatable characteristic of a population and was not an artifact
of random error or genotype by environment variances and
covariances. In general, the progeny variances for individual traits
from a particular year(s) were significantly correlated to the GVs
estimated in that year(s). the only exceptions being the 1990 yield
progeny variances with all GV estimates and the 1989-90 yield
progeny variances with the 1989 and 1990 GV estimates. An analysis
of the partial correlations between the progeny variances and the CNS
showed that the correlations of the yield progeny variances with the
CNS were lower when either the maturity or height progeny variance
was held constant. suggesting that yield progeny variance was probably
not important in determining the magnitude of the CV estimates.
Table 3 shows the correlations of the measurements of parental
genetic distance with the progeny variances for the individual traits.
All correlations possessed the expected sign under the hypothesis of
greater progeny variance with greater parental genetic distance. All
correlations were significant between the maturity progeny variances
of the populations and the CP, SI and PCD of the parents except for
those between the CP and the 1990 and 1989-90 maturity progeny
variances. The linear regressions of maturity progeny variance on the
OF. SI and PCD values were all significant (Table 4) though the R2
values were lower than the R2 values obtained by regressing the GVs
on these values indicating that the OF, SI, and PCD values were less
predictive of the maturity progeny variance of a population than of the
GV of a population. The 1989 and 1990 progeny variances contains
genetic and genotype by environment variance while the 1989-90
estimate contains only the genetic variance of a population. The
significant regressions of the genetic variance on all three measures of
parental genetic distance showed that they all were able to predict to
some degree the genetic variation for maturity of the populations. The
PCD measure of parental genetic distance was able to account for more
of the variability among the maturity genetic variances of the
populations than either the GP or $1 could and Figure 4 shows a plot of
the regression of the genetic variances of the populations on the PCD
values. The maturity of the parents of each population was determined
in 1990 and the difference between the parents of a population was

85

 

 

 

 

20 I I I I

>.
U
....|
S

u 15-
d
S
H
O
‘H

a, 10-
0
‘3
d
.,.|
H
g

5.—
0
«4
U
0
r:
Q
(9

o
1

 

2CD

Figure 4. Results of the linear regression of the maturity genetic
variances of the segregating populations on the principal component
distance (PCD) between the parents of the populations.

86

not correlated to the amount of progeny variance for maturity in a
population indicating that the magnitude of these variances were not
influenced by the phenotypic differences of the parents.

All three parental genetic distance measures were significantly
correlated to the progeny variances for plant height (Table 3). All
linear regressions of the height progeny variances of the populations
on either the OF, SI or PCD measures of parental distance were
significant but resulted in lower R2 values than the regressions of the
GVs of the populations on these measures, indicating that they had
less value in predicting the height progeny variance than the GV of a
population (Table 4). The CP and PCD measures had about equal value
in predicting the genetic variance for height of a population and were
able to account for more of the variation among these variances than
the SI measure could. Figure 5 shows the regression of the height
genetic variances of the populations on the PCD of the parents of the
populations. The height the parents of each population was
determined in 1990 and the difference between the parents of a
particular population was significantly correlated to the amount of
progeny and genetic variance for height in that population suggesting
that the larger phenotypic difi'erences of some of the parents may have
contributed to the larger height progeny variance of those populations.

While all three parental genetic distance measures were
significantly correlated to the 1989 yield progeny variances of the
populations, this was not repeated with the 1990 or 1989-90 data
(Table 3). The regression analyses produced similar results (Table 4)
(Figures 6 and 7). The R2 values of the genetic variance for yield with
all three measures were low. indicating that none were predictive of
this parameter. Table 5 summarizes the average yield progeny
variances for the populations whose parents had either a higher or a
lower OF. SI or PCD value than the mean of these measures. This
summary suggests that while none of the measures appeared to be
predictive of yield progeny variance and genetic variance. that in
general, populations that were derived from genetically distant
parents higher yield progeny variances and genetic variances than the
populations derived from parents that were similar. The yield
difference of the parents of each population was determined in 1990

87

 

 

 

 

15 I I I I
I

4)

:3

UI

.fi

0

:r:

H 10 -
O

‘H

0

U

G

d

or!

2

> 5

0

vi

4.)

0

£2

0

U

0
1

 

PCD

Figure 5. Results of the linear regression of the height genetic
variances of the segregating populations on the principal component
distance (PCD) between the parents of the populations.

88

 

 

 

 

1750 . . . .
I
'U
H
.3
s 1400
'U
0
0
U)
h 1050
O
'H
0
2 700
d
.fi
84
d
>
o 350
«4
U
0
i3
0
U o

 

PCD

Figure 6. Results of the linear regression of the seed yield genetic
variances of the segregating populations on the principal component
distance (PCD) between the parents of the populations.

89

 

 

 

7900
0
2 6410 - -
fl
”.1
H
d
:>
>. 4920 - -
i:
Q)
U!
0
H
m 3430 - -
'U
H
0
or.
" 1940 — -
Oi
Q
0‘
H

450 -

 

 

PCD

Figure 7. Results of the linear regression of the 1989 estimates of the
seed yield progeny variances of the segregating populations on the
principal component distance (PCD) between the parents of the

populations.

90

Table 5. The average seed yield progeny variances of the populations
whose parents had coefficient of parentage (CP). similarity index ($1)
or principal component distance (PCD) values that were either higher
(High) or lower (Low) than the mean of these measures, summarized
by years.

 

 

 

Distance High Year is)
measure or low 1989 1990 1989-90a
CP Low 4529.2 1092.0 774.3
C? High 1018.9 503.2 309.3
SI Low 4555.0 928.3 661.7
SI High 997.4 857.1 403.2
PCD Low 1991.1 838.6 490.7
PCD High 4276.8 975.6 600.6

 

a The 1989-90 progeny variance estimates the genetic variance for
seed yield within a population

and found not to be correlated to the amount of progeny variance for
yield.

Discussion

The amount of genetic variation in a segregating population
should increase as the genetic distance of the parents of the
population increases. The CP and SI estimate the percentage of the
loci that will be homozygous in a hybrid of two parents. As the CP and
SI values decrease, F1 heterozygosity should increase and the progeny
populations derived by selfing the F1 should become more variable
under the assumption of equal genetic effects at all loci. The PCD
measures the genetic distance between individuals and the
heterozygosity of an F1 should increase as the PCD of the parents
increase. All three measures of parental genetic distance were
significantly associated with the amount of overall variability in the
segregating populations analyzed in this study, suggesting that each

91

was predictive, to some degree, of the true genetic distance between
the parents. Each measure was able to account for approximately 60%
of the variation among the GVs of the populations estimated from the
1989-90 data set (Table 4) and therefore each appeared to be a useful
tool in selecting elite soybean parents that would produce variable
progeny populations from which a breeder could select diverse
phenotypes. Cowen and Frey (1987) found a similar association
between the C? of the parents and the GV of segregating oat
populations.

The ability of the measures of parental genetic distance to
predict the amount of variation among the F2 or F4 derived progeny of
a cross for individual traits was lower than their ability to predict the
GV of a population (Fable 4). This may be explained by the fact that all
three measures estimate the genetic distance between lines, averaged
across the whole genome. As such they should be more predictive of
the genetic distance for traits that are controlled by many loci
dispersed throughout the genome than for traits that are controlled by
relatively few loci as the violations of the assumptions of each measure
that occur at individual loci would tend to cancel each other out when
averaged over a larger number of loci. For example the CP assumes
that each inbred parent contributes equally to each locus in all
progeny of a cross when in reality an inbred progeny carries an allele
from only one of the parents. The discrepancies at individual loci will
average out as more loci are "sampled" by a trait such this assumption
will in effect be true when averaged across many loci and when the
other CP assumptions are not violated. The GV of a segregating
population is affected by the genetic distance of the parents at the loci
that control all of the measured traits such that it effectively samples
more loci than the individual traits and therefore the measures of
genetic distance may be more accurate for the loci influencing the GV
than for the loci contributing to the progeny variance or the genetic
variance of the individual traits.

None of the measures of parental genetic distance had a
repeatable association with the seed yield progeny variances and all
were unable to predict the seed yield genetic variance of a population
that results from crossing two elite soybean lines (Fable 4). While this

92

may limit the utility of these measures in a soybean cultivar
development program. the data presented in Table 5 and Figure 6 do
suggest some advantage in using these measures to decide which elite
lines to cross to produce segregating populations which will have a
higher likelihood of having an above average genetic variance for seed
yield. The measures of parental genetic distance were significantly
better at predicting the maturity and height progeny variance and
genetic variance of a population. though their ability to predict these
parameters were still relatively low (Table 4) suggesting that their
utility in a breeding program for these traits would be similar to that
described above for seed yield. The progeny variances estimated in a
single year contain genotype by environment variance that the
measures of parental genetic distance can not predict. Yet the CP, SI
and PCD values were still significantly associated with these estimate
of progeny variance for the traits maturity and height and this
association was repeated in both years of the study even when
additional populations were evaluated in 1990. This suggests that the
genotype by environment component of the single year progeny
variances for these traits was small or proportional to the genetic
variance. Seed yield generally has the lowest heritability of all the
measured traits (Brim. 1973) and it is possible that the single year
yield progeny variances contained proportionately larger contributions
of random or population specific genotype by environment effects than
the other traits. This may explain why the significant association of
the measures of parental distance with the 1989 estimates could not
be repeated in 1990. The low heritability of seed yield also makes it
harder to obtain an accurate estimate of the genetic variance of a
population and may explain why the measures of parental genetic
distance were unable to predict the magnitude of this parameter.

It is not clear which measure of parental genetic distance was
the better predictor of the (NS and the progeny or genetic variance of
a population as all were highly correlated to one another and produced
fairly similar results. The PCD distance was able to account for more
of the variability among the 1989-90 CNS and the genetic variances of
the individual traits of the populations than either the GP or $1
measures could (Table 4) and on this basis it appeared to be the better

93

measure. The PCD measure has the advantage of combining the purely
theoretical CP distance with actual data that evaluates genetic
distances at the DNA level and this may allow the shortcomings of one
measure to offset the shortcomings of the other. Cox et al. (1985)
noted that the biases of the CP and SI measures would tend to cancel
each other out. The SI data can monitor and adjust for violations of
the CP assumptions that all ancestral parents are completely unrelated
and that each parent contributes equally to all progeny while the CP
can estimate the probability of genetic distance at loci that were not
effectively assayed with linked ML. The PCD estimate of genetic
distance can be improved by evaluating more ML that can assay for
genetic differences in additional regions of the genome, while the CP
estimate of genetic distance can not be improved. The research of Lee
et al. (1989) indicates that the predictive value of a ML measure of
genetic distance for a particular trait may be improved by assaying for
genetic differences only in those regions of the genome that affect the
trait of interest thereby eliminating useless contributions to the
genetic distance estimate from irrelevant portions of the genome.
Pertinent chromosome segments could be identified through
ML/ quantitative trait loci studies using mapped ML.

In conclusion. all three measures of parental genetic distance
appeared to have utility in selecting elite soybean parents that when
crossed would have a high likelihood of producing segregating
populations with above average genetic variability. The use of such
measures could allow a breeder to eliminate crossing certain parents
and allow cultivar development resources to be allocated to
populations with a greater potential of producing outstanding progeny.

List of References

 

List ofReferenees

Brim. C.A.. 1973. Quantitative genetics and breeding: in Soybeans: Improvement.
Production and Uses. ed B. Caldwell. American Society of Agronomy. Madison.

Buttery. BR. and R1. Buzzel]. 1968. Peroxidase activity in seeds of soybean varieties.
Crop Sci. 8:722-725.

Cardy. E..]. and “ID. Beversdorf. 1984. A procedure for starch gel electrophoretic
detection of isozymes of soybean (Glycine max L. Merr.). Departemnt of Crop Science
Technical Bulletin 1 19/ 8401. University of Guelph. Guelph Ontario. Canada. N1G 2W1.

Cowen. N.M.. and K.J. Frey. 1987. Relationship between genealogical distance and
breeding behavior in oats. Euphytica 36:413-424.

Cox. T.S.. Y.T. Kiang. M.B. Gorman. and D.M. Rodgers. 1985. Relationship between
coefiicient of parentage and genetic similarity indices in the soybean. Crop Sci. 25:529-
532.

Cox. T.S.. and J.P. Murphy. 1990. The effect of parental divergence on F3 heterosis in
winter wheat crosses. Theor. Appl. Genet. 79:241-250.

Doong. J .Y.H.. and Y.T. Kiang. 1987. Cultivar identification by isozyme analysis. Soy.
Genet. Newsl. 14:189-225.

Falconer. D.S.. 1981. Introduction to quantitative geneiim. Longrnan House. New York.

Fehr. “IR. and LB Ortiz. 1975. Registration of a soybean germplasm population. Crop
Sci 15:739.

Frei. O.M.. C.W. Stuber. and MM. Goodman. 1986. Use of allozymes as genetic markers
for predicting performance in maize single cross hybrids. Crop Sci. 26:37-42.

Godshalk. R.B.. M. Lee. and KR. Lamkey. 1990. Relationship of restriction fragment
length polymorphisms to single-cross hybrid performance of maize. Theor. Appl.
Genet. 80:273-280.

Goodman. MM. 1968. A measure of ”overall variability“ in populations. Biometrics
24:189-192.

Goodman. MM. 1972. Distance analysis in biology. Syst. Zool. 21:174-186.

Gorman. M.M.. Y.T. Kiang. Y.C. Chiang. and RC. Palmer. 1983. Electrophoretic
classification of the early maturity groups of named soybean cultivars. Soybean Genet.
Newsl. 9:143-156.

Hadjinov. M.l.. V.S. Scherbak. N.l. Benko. V.P. Gusev. T.B. Sukhorzhevskaya. and L.P.

Voronova. 1982. interrelationship between isozymic diversity and combining ability
in maize lines. Maydica 27:135-149.

94

95

Hunter. R.B.. and L.W. Kannenberg. 1971. Isozyme characterization of corn (Zea mays)
inbreds and it's relationship to single cross hybrid performance. Can. J. Gent. Cytol.
13:649-655.

Kempthorne. 0.. 1969. An Introduction to Genetic Statistics. lowa State University
Press. Ames.

Kiem. P.. R.C. Schumaker. and RG. Palmer. 1989. Restriction fragment length
polymorphism diversity in soybean. Theor. Appl. Genet. 77:786-792.

Lamkey. K.R.. A.R. Hallauer. and A.L. Kahler. 1987. Allelic differences at enzyme loci
and hybrid performance in maize. J. Hered. 78:231-234.

Lee. M.. E.B. Godshalk. K.R. Lamkey. and W.W. Woodman. 1989. Association of
restriction fragment length polymorphisms among maize inbreds with agronomic
performance of their crosses. Crp Sci. 29:1067-1071.

Lefort-Buson, M.. Y. Datee. and B. Guillot-Lemoine. 1987. Heterosis and genetic
distance in rapeseed (Brassioa napus L): use ofkinship coefficient Genome 29:11-18.

Sneller. C.. 1991. Estimation of the genetic structure of an elite soybean population and
it‘s application to a soybean breeding program. Ph D dissertation. Michigan State
University.

Sokal. RR. 1965. Statistical methods in systematics. Bio. Rev. Camb. Phil. Soc.
40:337-341.

Sokal.R.R.. and P.H.A. Sneath. 1963. Principles of Numerical Taxonomy. Freeman.
San Francisco.