Code & data – Understanding factors influencing the estimated genetic variance and the distribution of breeding values
The attached files are different formats of the same Jupyter Notebook containing R code. This document is about finding distribution properties of the genetic relationship matrix (K) and its inverse (used in best linear unbiased prediction (BLUP)) influencing the genetic variance, the estimated heritability (h^2), and the variance of estimated breeding values. The main factors influencing the distribution of estimated breeding values are the phenotypic variance, estimated h^2, and K^-1. Matrices K and K^-1 underwent transformations to find the distribution properties of these two matrices affecting the estimated h^2 and the distribution of estimated breeding values. The wheat data from R package BGLR were used. The inverses of the pedigree-based K (A^-1) and the genomic-based K (G^-1) had similar averages of diagonal and similar averages of off-diagonal elements but very different h^2 estimates. Therefore, the distributions of A^-1 and G^-1 elements were studied closely.
Steps to reproduce
Steps to reproduce are presented in the attached files, which are different formats of the same Jupyter Notebook containing R code.