Making Whole Genome Predictions In The World Of Cheap Genotyping and Difficult Phenotyping: Implications For Plant Breeding In The Developing World

In plant breeding we are facing a world of increasingly inexpensive genotyping.  Whole genome profiles can now be routinely obtained on every line in a breeding program.  Phenotyping on the other hand, remains time consuming, difficult and expensive.  Application of genomic selection to make whole-genome predictions of the breeding value of selection candidates prior to phentoyping shows great promise for increasing the rate of gain in plant breeding programs by enabling selection on larger populations prior to phentoyping evaluation.  The development of these models, however, relies on a training population that has both high quality phenotypes and dense genome-wide molecular markers. This phenotypic data is not easy or cheep to obtain. High-quality phenotypic data on large populations is a prerequisite to applying genomics-assisted breeding. We are therefore looking at simple, but innovative ways to improve both the collection and management of phenotypic in applied breeding programs.