Genome-wide effects of artificial selection

Humans have been artificially selecting for favorable traits in crops, pets, and livestock over millennia. Years of theoretical predictions and experimental evolution studies have shown the detrimental effects of increased homozygosity, and the population-wide advantages of artificially maintaining heterozygosity. Two new manuscripts (Hedrick (2015), and Kessner and Novembre (2015)) aim at discussing the genome wide effects of artificial selection, albeit focused on different characteristics.

Expected increase in heterozygosity upon introduction of a new allele at a low frequency, plotted at different levels of artificial selection. Image courtesy: http://jhered.oxfordjournals.org/content/106/2/141/F2.expansion.html

Hedrick (2015) reviews the phenomenon of ‘overdominance’, or heterozygote advantage across 12 different traits. The list includes milk yield in dairy cattle, litter size in pigs, tail length in cats, hairlessness in dogs, among other described cases. Hedrick also discusses some theoretical predictions on the perpetration of heterozygosity in artificially selected populations, importantly the strength of selection on fixation times.

…given very strong selection from an environmental or other change, a new mutant that has an advantage as a heterozygote might increase in frequency. However, if it had a lowered fitness as a homozygote, it would be maintained as a polymorphism due to its overall heterozygote advantage…

Kessner and Novembre (2015) describe an analysis pipeline using forqs, and utilize simulations to predict the effects of artificial selection at QTLs, based on population sizes (drift, efficacy of selection), proportion of individuals chosen to propagate, length of the experiment, and replication. Forward simulations indicate (1) predictable trajectories of fixation of new alleles under constant selection (also predicted by the scenarios described by Hedrick (2015) above), (2) qualitative differences in trajectories of allele frequencies of linked QTLs, (3) increase in power to detect selected QTLs (versus neutrally evolving loci) with increased replication, and starting population sizes, and importantly, (4) the effect of recombination (which reduces the interference, and LD between linked QTLs).

We emphasize that the opportunity for recombination is a key factor in the power to detect and localize QTLs, and that this should be taken into account by future designers of artificial selection experiments.

References:

Hedrick, Philip W. “Heterozygote Advantage: The Effect of Artificial Selection in Livestock and Pets” J Hered (2015) 106 (2): 141-154 http://dx.doi.org/10.1093/jhered/esu070

Kessner, Darren, and John Novembre. “Power analysis of artificial selection experiments using efficient whole genome simulation of quantitative traits.” Early Online February 10, 2015, http://dx.doi.org/10.1534/genetics.115.175075

Kessner, Darren, and John Novembre. “forqs: forward-in-time simulation of recombination, quantitative traits and selection.” Bioinformatics 30.4 (2014): 576-577. http://dx.doi.org/10.1093/bioinformatics/btt712

Share

About Arun Sethuraman

I am a computational biologist, and I build statistical models and tools for population genetics. I am particularly interested in studying the dynamics of structured populations, genetic admixture, and ancestral demography.
This entry was posted in bioinformatics, evolution, genomics, methods, mutation, population genetics, theory. Bookmark the permalink.