Genetics of complex traits: prediction of phenotype, identification of causal polymorphisms and genetic architecture

Abstract

Complex or quantitative traits are important in medicine, agriculture and evolution, yet, until recently, few of the polymorphisms that cause variation in these traits were known. Genome-wide association studies (GWAS), based on the ability to assay thousands of single nucleotide polymorphisms (SNPs), have revolutionized our understanding of the genetics of complex traits. We advocate the analysis of GWAS data by a statistical method that fits all SNP effects simultaneously, assuming that these effects are drawn from a prior distribution. We illustrate how this method can be used to predict future phenotypes, to map and identify the causal mutations, and to study the genetic architecture of complex traits. The genetic architecture of complex traits is even more complex than previously thought: in almost every trait studied there are thousands of polymorphisms that explain genetic variation. Methods of predicting future phenotypes, collectively known as genomic selection or genomic prediction, have been widely adopted in livestock and crop breeding, leading to increased rates of genetic improvement.

Keywords: complex traits; genome-wide association studies; genomic selection.

Figure 1.

Genome-wide analysis of bovine milk fat percentage showing results for a region around the FASN gene. (a) Posterior probability that an SNP has a non-zero effect from Bayes R where all SNPs are fitted in the model simultaneously. (b) −log₁₀ p-value from GWAS single SNP regressions. The top Bayes R variant is annotated (with base pair position) and shown as a purple diamond, and the strength of LD (r²) between this and all other variants is colour coded. (Adapted from data published in [20].)

Figure 2.

Local GEBV variance near the PAEP gene for FY, MY and PY using Bayes R and BLUP. Shown is the GEBV variance in overlapping 250 kb windows for Holstein and Jersey reference animals from SNP effects estimated from the multi-breed reference population. Traits: FY, fat yield; MY, milk yield; PY, protein yield. The position of PAEP on BTA11 is marked (*). Note the changed y-axis scale for each graph. Adapted from Kemper et al. [14]. (Online version in colour.)

Figure 3.

Comparison of results for three genome-wide analysis methods of bovine milk protein percentage in a region around the LALBA gene. (a) Bayes RC incorporates prior biological information on sites that are more likely to have an effect on the trait, while (b) Bayes R and (c) GWAS assume all sites are equally likely to affect the trait. The top Bayes RC variant is shown in purple (with base pair position), and the strength of LD (r²) between this and all other variants is colour coded (adapted from data published in [20]).