468
Predictive Performance Yielded by Approximate Genomic Matrices

Monday, August 18, 2014
Posters (The Westin Bayshore)
Eduardo Manfredi , INRA, Toulouse, France
Clement Carre , IMT Université Paul Sabatier, Toulouse, France
Miguel A Toro , Animal Production Department, Universidad Politecnica de Madrid, Madrid, Spain
Abstract Text:

Some methods for genetic prediction require the inverse of large genomic variance-covariance matrices amongst individuals. We studied by simulation the predictive performance of approximate inverses of genomic matrices, relatively to the usual complete genomic matrix. The approximate inverses were easy to compute but accuracies dropped by up to 30% according to the studied genetic scenarios.

Keywords:

genetic prediction, genomic selection, genomic matrices