GBS Analysis of Selection Footprints in Alfalfa (Medicago sativa) Populations Recurrently Selected for Freezing Tolerance

Insufficient tolerance to freezing temperatures reduces winter survival of alfalfa (Medicago sativa) in northern climates. The identification of allelic variants associated with variation in freezing tolerance within breeding populations would help accelerate the improvement of complex traits in open-pollinated polyploids.

We used GBS to identify SNPs under selection pressure in populations of outbreeding alfalfa recurrently selected for superior freezing tolerance (FT) within the cultivar Apica. GBS was performed on ApeKI-digested libraries using DNA from 48 genotypes of each of two populations: A-FT0 (initial population) and A-FT5 (obtained after 5 cycles of recurrent selection). Nearly 400 million 100-bp reads were obtained on an Illumina HiSeq2000 and analyzed using the UNEAK pipeline. After filtering for quality and read depth, 2732 high-quality SNPs were retained, of which 1467 could be positioned on the Medicago truncatula reference genome (Mt4.0).

Association between SNP loci (pairwise Fisher exact test with FDR correction) revealed that more than 90% (2531) have no significant linkage. A Fisher exact test identified 103 SNP loci displaying significant (p-value ≤0.01) genotype frequency variation in response to recurrent selection. Among this group, 73 had homologies with M. truncatula and/or M. sativa coding sequences. Several of these sequences are putatively involved in signal transduction and regulation of transcription and translation. The candidate loci located on M. truncatula were distributed along the eight chromosomes with no apparent clustering.

Results show that GBS analysis of populations under recurrent selection allows the identification of regions with candidate genes putatively affecting quantitative traits in complex polyploids.