Large-Scale Discovery of Gene-Enriched SNPs

Whole-genome association studies of complex traits in higher eukaryotes require a high density of single nucleotide polymorphism (SNP) markers at genome-wide coverage. To design high-throughput, multiplexed SNP genotyping assays, researchers must first discover large numbers of SNPs by extensively resequencing multiple individuals or lines. For SNP discovery approaches using short read-lengths that next-generation DNA sequencing technologies offer, the highly repetitive and duplicated nature of large plant genomes presents additional challenges. Here, we describe a genomic library construction procedure that facilitates pyrosequencing of genic and low-copy regions in plant genomes, and a customized computational pipeline to analyze and assemble short reads (100-200 bp), identify allelic reference sequence comparisons, and call SNPs with a high degree of accuracy. With maize (Zea mays L.) as the test organism in a pilot experiment, the implementation of these methods resulted in the identification of 126,683 putative SNPs between two maize inbred lines at an estimated false discovery rate (FDR) of 15.1%. We estimated rates of false SNP discovery using an internal control, and we validated these FDR rates with an external SNP dataset that was generated using locus-specific PCR amplification and Sanger sequencing. These results show that this approach has wide applicability for efficiently and accurately detecting gene-enriched SNPs in large, complex plant genomes.