Retention of agronomically important variation in germplasm core collections: implications for allele mining

Authors: Reeves, Patrick; Panella, Leonard; Richards, Christopher

Submitted to: Theoretical and Applied Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/16/2011
Publication Date: 1/7/2012
Citation: Reeves, P.A., Panella, L.W., Richards, C.M. 2012. Retention of agronomically important variation in germplasm core collections: implications for allele mining. Theoretical and Applied Genetics. 124(6) 1155-1171. DOI: 10.1007/S00122-011-1776-4.

Interpretive Summary: Core collections—highly diverse, reduced representations of complex germplasm collections—will be necessary to expedite discovery of new, agronomically useful traits in wild germplasm. It is becoming increasingly popular to use evolutionarily neutral genetic marker loci to help determine which accessions belong in a core collection. We show that genetic diversity at genes of agronomic importance, because they have evolved under natural or artificial selection, will not likely be distributed in parallel with diversity at neutral loci. Therefore, core collections assembled using neutral genetic loci as a guide are not likely to contain elevated allelic diversity at genes of agronomic interest. Core collections based on neutral markers alone will not be useful for allele mining, which seeks to identify valuable traits by directly assaying molecular variation at agronomic genes. Detailed understanding of the evolutionary history of a collection may aid in identification of accessions with elevated diversity at genes of interest.

Technical Abstract: The primary targets of allele mining efforts are loci of agronomic importance. Agronomic loci typically exhibit patterns of allelic diversity consistent with a history of natural or artificial selection. Natural or artificial selection causes the distribution of genetic diversity at such loci to deviate substantially from the pattern found at neutral loci. The germplasm utilized for allele mining should contain maximum allelic variation at loci of interest, in the smallest possible number of samples. We show that the popular core collection assembly procedure “M” (marker allele richness), which leverages variation at neutral loci, performs worse than random assembly for retaining variation at a locus of agronomic importance in sugar beet (Beta vulgaris L. subsp. vulgaris) that is under selection. We present a corrected procedure (“M+”) that outperforms M. An extensive coalescent simulation was performed to demonstrate more generally the retention of neutral versus selected allelic variation in core subsets assembled with M+. A negative correlation in level of allelic diversity between neutral and selected loci was observed in 42% of simulated data sets. When core collection assembly is guided by neutral marker loci, as is the current common practice, enhanced allelic variation at agronomically important loci should not necessarily be expected.