Beyond Genomic Selection

Abstract Text:

Use of genomic information to select for identified QTL has essentially failed, working only for traits affected by one or a few QTL.  We resort to additive statistical fits and empirical extrapolation, with no inference of mechanism let alone identification of QTL.  Although this works well for many scenarios, we could do better, especially for traits involving high levels of non-additive variation.    However, even in these cases we can probably only work with traits of few QTL.  Biologically inspired mechanistic models may compete favourably with statistical models involving epistatic parameters.  There is an emerging need for systems to ensure ongoing phenotyping for genomic calibration, and proposals for this are discussed.  The development of in-vitro reproductive technologies would greatly increase the impact of genomic information, with selection between zygotes and even between gametes providing new levels of genetic information and gain.

Keywords:

Genomic Selection

Epistasis

Reproductive technologies