Comparison Between Core Set Selection Methods Using Different Illumina Marker Platforms: A Case Study of Assessment of Diversity in Wheat

Collections of plant genetic resources stored in genebanks are an important source of genetic diversity for improvement in plant breeding programs and for conservation of natural variation. The establishment of reduced representative collections from a large set of genotypes is a valuable tool that provides cost-effective access to the diversity present in the whole set. Software like Core Hunter 3 is available to generate high quality core sets. In addition, general clustering approaches, e.g., k-medoids, are available to subdivide a large data set into small groups with maximum genetic diversity between groups. Illumina genotyping platforms are a very efficient tool for the assessment of genetic diversity of plant genetic resources. The accumulation of genotyping data over time using commercial genotyping platforms raises the question of how such huge amount of information can be efficiently used for creating core collections. In the present study, after developing a 15K wheat Infinium array with 12,908 SNPs and genotyping a set of 479 hexaploid winter wheat lines (Triticum aestivum), a larger data set was created by merging 411 lines previously genotyped with the 90K iSelect array. Overlaying the markers from the 15K and 90K arrays enabled the identification of a common set of 12,806 markers, suggesting that the 15K array is a valuable and cost-effective resource for plant breeding programs. Finally, we selected genetically diverse core sets out of these 890 wheat genotypes derived from five collections based on the common markers from the 15K and 90K SNP arrays. Two different approaches, k-medoids and Core Hunter 3 were compared,and k-medoids was identified as an efficient method for selecting small core sets out of a large collection of genotypes while retaining the genetic diversity of the original population.