Classification and Characterization of Species within the Genus Lens Using Genotyping-by-Sequencing (GBS)

Lentil (Lens culinaris ssp. culinaris) is a nutritious and affordable pulse with an ancient crop domestication history. The genus Lens consists of seven taxa, however, there are many discrepancies in the taxon and gene pool classification of lentil and its wild relatives. Due to the narrow genetic basis of cultivated lentil, there is a need towards better understanding of the relationships amongst wild germplasm to assist introgression of favourable genes into lentil breeding programs. Genotyping-by-sequencing (GBS) is an easy and affordable method that allows multiplexing of up to 384 samples or more per library to generate genome-wide single nucleotide Polymorphism (SNP) markers. In this study, we aimed to characterize our lentil germplasm collection using a two-enzyme GBS approach. We constructed two 96-plex GBS libraries with a total of 60 accessions where some accessions had several samples and each sample was sequenced in two technical replicates. We developed an automated GBS pipeline and detected a total of 266,356 genome-wide SNPs. After filtering low quality and redundant SNPs based on haplotype information, we constructed a maximum-likelihood tree using 5,389 SNPs. The phylogenetic tree grouped the germplasm collection into their respective taxa with strong support. Based on phylogenetic tree and STRUCTURE analysis, we identified four gene pools, namely L. culinaris/L. orientalis/L. tomentosus, L. lamottei/L. odemensis, L. ervoides and L. nigricans which form primary, secondary, tertiary and quaternary gene pools, respectively. We discovered sequencing bias problems likely due to DNA quality and observed severe run-to-run variation in the wild lentils. We examined the authenticity of the germplasm collection and identified 17% misclassified samples. Our study demonstrated that GBS is a promising and affordable tool for screening by plant breeders interested in crop wild relatives.

小扁豆（Lens culinaris ssp. culinaris）是一种营养丰富且经济实惠的食用豆科作物，拥有悠久的作物驯化历史。兵豆属（Lens）共计7个分类群，但小扁豆及其野生近缘种的分类与基因库（gene pool）划分仍存在诸多争议。由于栽培小扁豆的遗传基础狭窄，亟需深入解析野生种质资源间的亲缘关系，以助力将优良基因渐渗至小扁豆育种项目中。
测序分型（Genotyping-by-Sequencing, GBS）是一种简便且经济的技术，可实现每个文库最多384份或更多样本的多重测序，从而获得全基因组范围的单核苷酸多态性（Single Nucleotide Polymorphism, SNP）标记。本研究采用双酶切GBS策略，对本实验室的小扁豆种质资源库进行分型鉴定。我们构建了2个96重混合GBS文库，共包含60份种质材料，其中部分种质设有多个重复样本，且每份样本均设置2次技术重复测序。
本研究开发了一套自动化GBS分析流程，共检测到266,356个全基因组SNP位点。基于单倍型信息过滤低质量与冗余SNP后，我们利用5,389个高质量SNP位点构建了最大似然树（maximum-likelihood tree）。系统发育树将供试种质资源按各自分类群清晰聚类，且聚类支持率极高。结合系统发育树与STRUCTURE群体结构分析结果，我们共鉴定出4个基因库，分别为L. culinaris/L. orientalis/L. tomentosus、L. lamottei/L. odemensis、L. ervoides以及L. nigricans，依次对应一级、二级、三级与四级基因库。
本研究发现了可能由DNA质量导致的测序偏差问题，且野生小扁豆样本存在显著的批次间测序差异。我们对种质资源库的材料真实性进行了核查，鉴定出17%的错分样本。本研究证实，GBS是一款极具应用前景且经济高效的筛查工具，可服务于关注作物野生近缘种的植物育种工作者。