Skip to main content

Archived Comments for: A class of Bayesian methods to combine large numbers of genotyped and non-genotyped animals for whole-genome analyses

Back to article

  1. Allelic frequencies and extra parameter "mu_g"

    Andres Legarra, INRA

    28 October 2014

    We would like to make a comment on the very complete work by Fernando et al. The issue of the effect of selection or drift on the allelic frequencies of genotyped animals with respect to the pedigree founders has been frequently addressed in the context of Single Step GBLUP. Unfortunately, some relevant works have not been cited in the present paper, and we consider that they deserve being mentioned in order to show readers the intimate connexion between different methods.

    The fact that allelic frequencies or, equivalently, breeding values drift from pedigree base populations to genotyped populations was addressed by Vitezica et al. (Genet. Res., Camb. (2011), 93, pp. 357–366) and Meuwissen et al. (J Anim Breed Genet 2011, 128:429-439 [38]). Vitezica et al. model includes an extra unknown "mu" (see their Appendix), which is strictly equivalent to the "mu_g" suggested by Fernando et al. in this work.

    Vitezica et al. also showed that "mu" is a random variable (i.e., drift is a stochastic process in a finite population) and therefore they could integrate mu out of the distribution of breeding values, resulting in a genomic relationship matrix that is adjusted towards pedigree-based relationship matrix.

    Similar ideas were developed in parallel by Meuwissen et al. (J Anim Breed Genet 2011, 128:429-439 [38]) and Powell et al. (Nature Reviews Genetics 11, 800-805, 2010), although the term "mu" does not appear explicitely in their derivations. The three papers (Powell et al., Meuwissen et al., Vitezica et al.) also consider the reduction in genetic variance from the base to the genotyped population.

    Later, Christensen (Genetics Selection Evolution 2012, 44:37) suggested yet another different method which uses markers (and not pedigree) as the reference base.

    Signed, Ole F Christensen and Andres Legarra

    Competing interests

    We do not have competing interests.