If Allele Mining Gets Difficult, Try Genomic Selection

Allele mining activities include 1. Assembly of germplasm with potentially interesting novel alleles and 2. Evaluation trials and genotyping of this germplasm to "dissect" traits and identify markers associated with valuable traits. Plans are then made to introgress those markers and hopefully the causal polymorphisms in linkage disequilibrium with the identified markers. These actions operate on few loci at a time and deal in suboptimal ways with the possibility of error in association identification. The purpose of the actions is to introduce novel variation into elite populations to improve the performance of varieties. A highly valuable side product, however, is the identification and potential cloning of causal genes. A potential alternative to allele mining to improve elite populations is genomic selection in a population admixed between elite and exotic germplasm. This approach does not seek to identify causal genes but uses all information available directly to estimate or predict the performance of introgression lines that can become prebreeding parents. This approach can be effective for improvement because it is optimized directly for the desired goal of improvement. Furthermore it works well on many, potentially dozens or hundreds of loci simultaneously. On the other hand, it is suboptimal for the purpose of identifying causal genes and mechanisms. In this presentation I will discuss approaches to use genomic selection with admixed elite / exotic populations.