Page 193 - Linear Models for the Prediction of Animal Breeding Values 3rd Edition

11 Computation of Genomic

Breeding Values and

Genomic Selection

11.1 Introduction

In outbreeding populations, the incorporation of molecular information in breeding
programmes on the basis of the linkage analysis, as discussed in Chapter 10, is limited,
as the marker maps are rather sparse and linkage between the markers and QTL may
not be sufficiently close enough to persist across the population. Thus the linkage
phase between marker and QTL must be established for every family in which the
marker is intended to be used for selection.
However, a huge amount of variation has been discovered in the genome at the
DNA level as a result of sequencing the genomes of most livestock species. The
most abundant form of variation is the single nucleotide polymorphisms (SNPs).
An SNP is a DNA sequence variation occurring when a single nucleotide (A, T, C
or G) in the genome differs between paired chromosomes in an individual. For
example, two sequenced DNA fragments from different individuals, AAGCCTA to
AAGCTTA, contain a difference in a single nucleotide. In this case we say that
there are two alleles: C and T. Generally, SNPs are diallelic. In view of the high
frequency of SNPs in the genome, and developments in genotyping technology that
mean many thousands of SNPs can be genotyped very cheaply, they have been
proposed as markers for use in QTL analysis and in association studies in place of
microsatellites.
The main emphasis of this chapter is the use of SNPs to directly compute EBVs
of animals, which are often called direct genomic breeding values (DGV). This is usu-
ally combined with some measure of the traditional EBV, say parent index, from an
animal model to produce what is termed genomic breeding values (GEBV), which are
officially published and used for the selection of animals.
The use of GEBV in the selection of animals has been referred to as genomic
selection. Genomic selection requires that markers (SNPs) are in linkage disequilib-
rium (LD) with the QTLs across the whole population. LD can be defined as the
non-random association between the alleles of two loci (e.g. between alleles of a
marker and a QTL). Given a marker locus, A (with alleles A , A ), and a QTL locus,
1 2
B (with alleles B and B ), on the same chromosome, LD can be measured as the
1 2
2
squared correlation (r ) between the marker and the QTL as:
D = freq(A B )*freq(A B ) − freq(A B )*freq(A B )
2 1
1 1
2 2
1 2
2
2
r = D /[freq(A )*freq(A )*freq(B )*freq(B )]
1 2 1 2
2
The r between the marker and the QTL indicates the proportion of the variance for
the QTL that can be explained at the marker.
© R.A. Mrode 2014. Linear Models for the Prediction of Animal Breeding Values, 177
3rd Edition (R.A. Mrode)