Find in Library

<p>Abstract</p> <p>Background</p> <p>The genetic diversity of crop species is the result of natural selection on the wild progenitor and human intervention by ancient and modern farmers and breeders. The genomes of modern cultivars, old cultivated landraces, ecotypes and wild relatives reflect the effects of these forces and provide insights into germplasm structural diversity, the geographical dimension to species diversity and the process of domestication of wild organisms. This issue is also of great practical importance for crop improvement because wild germplasm represents a rich potential source of useful under-exploited alleles or allele combinations. The aim of the present study was to analyse a major <it>Pisum </it>germplasm collection to gain a broad understanding of the diversity and evolution of <it>Pisum </it>and provide a new rational framework for designing germplasm core collections of the genus.</p> <p>Results</p> <p>3020 <it>Pisum </it>germplasm samples from the John Innes <it>Pisum </it>germplasm collection were genotyped for 45 retrotransposon based insertion polymorphism (RBIP) markers by the Tagged Array Marker (TAM) method. The data set was stored in a purpose-built Germinate relational database and analysed by both principal coordinate analysis and a nested application of the Structure program which yielded substantially similar but complementary views of the diversity of the genus <it>Pisum</it>. Structure revealed three Groups (1-3) corresponding approximately to landrace, cultivar and wild <it>Pisum </it>respectively, which were resolved by nested Structure analysis into 14 Sub-Groups, many of which correlate with taxonomic sub-divisions of <it>Pisum</it>, domestication related phenotypic traits and/or restricted geographical locations. Genetic distances calculated between these Sub-Groups are broadly supported by principal coordinate analysis and these, together with the trait and geographical data, were used to infer a detailed model for the domestication of <it>Pisum</it>.</p> <p>Conclusions</p> <p>These data provide a clear picture of the major distinct gene pools into which the genus <it>Pisum </it>is partitioned and their geographical distribution. The data strongly support the model of independent domestications for <it>P. sativum ssp abyssinicum </it>and <it>P. sativum</it>. The relationships between these two cultivated germplasms and the various sub-divisions of wild <it>Pisum </it>have been clarified and the most likely ancestral wild gene pools for domesticated <it>P. sativum </it>identified. Lastly, this study provides a framework for defining global <it>Pisum </it>germplasm which will be useful for designing core collections.</p>