Mining Plant Collections Using Phenotyping and High-Throughput Sequencing for Accelerated Crop Improvement - a Case Study of the Native American Bean Apios americana

The evaluation of plant collections through combination of phenotyping and genomics provides a valuable tool for crop improvement. A study involving multi-year, multi-environment phenotypic evaluation, generation of genomic resources and their utilization in mining a collection of Apios americana is presented. Apios is native to central and eastern USA and was once a staple food for Native Americans. It produces protein rich tubers (~16%; 3X to 4X that of Solanum tuberosum) and is cultivated in Japan and South Korea because of its nutritional benefits. The phenotypic evaluation of 53 Apios accessions utilizing 20 traits derived from above and belowground growth parameters was conducted in Iowa, Virginia, and Pennsylvania with lines that produced a range of tuber yields up to 1,515 g per plant. We have generated a reference transcriptome assembly and gene expression atlas, and have genotyped the entire breeding collection using RNA-Seq to produce ~58,000 SNPs scorable across the collection. Population structure analysis identifies six subgroups in the collection. To estimate linkage disequilibrium (LD), we mapped Apios SNPs onto the Phaseolus vulgaris chromosomes (giving an average of ~4,200 markers per chromosome). LD decays rapidly beyond 10-15 kb. The high density of SNPs allowed partial pedigree re-construction, which is generally consistent with historical records. Association analysis using SNPs and gene-expression markers identified promising parents, marker-trait associations and candidate genes for 14 phenotypic traits. A first set of crosses based on these results are at the first filial generation. This study demonstrates the possibility of rapidly improving under-studied and novel crops.