Plants are under threat from bacteria, viruses, fungi, and nematodes in their environment. Plants deploy a sophisticated network of defense responses to avoid and defeat these pathogens, and in response pathogens counteract plant defenses through their own suite of biochemical weapons and signaling molecules. These battles are pervasive in both natural settings and agriculture. The goal of this dissertation research is to provide insight into genetic variation present in diverse populations of Beta vulgaris and identify patterns of disease resistance (R) gene variation.Nucleotide-binding (NB-ARC), leucine-rich-repeat genes (NLRs) account for 60.8% of R genes molecularly characterized from plants. NLRs exist as large gene families prone to tandem duplication and transposition, with high sequence diversity among crops and their wild relatives. I used the conserved NB-ARC domain to build a B.vulgaris-specific hidden Markov model (HMM). The HMM identified 231 tentative NB-ARC loci in a highly contiguous genome assembly of sugar beet, revealing diverged and truncated NB-ARC signatures as well as full-length sequences. The putative NB-ARC-associated proteins contained NLR resistance gene domains, including Toll/interleukin-1 receptor (TIR), coiled-coil (CC), and leucine-rich repeat (LRR), as well as other integrated domains. HMM-based domain detection was extended to 23 populations encompassing four crop types of B. vulgaris. Whole-genome sequences were generated by pooling 25 individuals per population, then sequencing each population in a single bulk reaction using 2x150 bp chemistry. These reads were assembled de novo to efficiently capture population-wide genetic variation. The nucleic-acid-based NB-ARC HMM was used to scan de novo contigs and infer genetic variation within and between populations, which identified an average of 139.5 NB-ARC domains per population. The pooled population sequencing strategy was expanded to 71 populations total. Short reads were used in a targeted reassembly pipeline to detect structural variation in each of the 71 populations. This method identified 4,995,443 indels with lengths under 1 kb. These indels were analyzed for chromosome position, length in bp, and frequency across populations, and revealed non-random patterns of indel variation. Half of the indels were detected in five or more populations, suggesting that indel assembly from pooled population sequences is reproducible. Furthermore, indels were sufficient to differentiate populations by crop type, supporting the conclusion that the data modeled genetic differences originating in historical crop development. Divergence in the population-wide distribution of seven- and eight-bp indels led to identification of an enriched sequence motif, suggesting possible biological function of the sequence such as a TE target site duplication or transcription factor binding site.This work presents the first detailed view of NLR family composition in a member of the Caryophyllales and demonstrates an additional nucleic-acid-based method for resistance gene prediction in non-model plant species. Pooled population sequencing was used to access novel variation in breeding populations of B. vulgaris and identify structural variants that reflected underlying genotypic relationships. Future work will build on resistance gene modeling, pooled population sequencing, and detection of genetic variation to aid breeding for disease resistance.
Read
