Selection keeps an extra-close eye on multi-functional genes

Victorinox Swiss Army Knife

More functions, more selective constraint? Photo by James Case.

Genes that have roles in multiple traits—pleiotropic genes—have long been thought to be under stronger selection as a result of those multiple functions. The basic logic is that, when a gene produces a protein that has a lot of different functional roles, there are more functions that will be disrupted by changes to that protein. Which would be more inconvenient: if your smartphone suddenly needed a new type of power connector, or if every electrical outlet in your house suddenly accepted only plugs with four prongs?

As much sense as that makes, we don’t have a lot of direct evidence that pleiotropic genes experience stronger selection. A paper just released online ahead of print at Genetics provides just that, using the suite of genetic resources available for the subject of some of the original experiments in evolutionary genetics, Drosophila.

Katrina McGuigan and coauthors at the University of Queensland started from a conclusion of quantitative genetics theory, that the strength of selection acting against mutations in a quantitative trait is equal to the ratio of the variation introduced by new mutations to the standing genetic variation in that trait.

To estimate these two statistics, they used two panels of Drosophila serrata inbred lines: one panel of lines (MA) raised from a single inbred ancestral line over more than two dozen generations with small effective population size to help accumulate mutations; the other panel (G) consisting of lines sampled from a wild population. Trait variation among the MA lines is due to just that—accumulation of mutations since they diverged from their common ancestor. On the other hand, trait variation among the G lines should reflect genetic variation in the wild population from which they were descended. Thus, estimates of mutational variation, and standing genetic variation.

The “traits” McGuigan et al. examined in these two panels of flies were rates of gene expression assayed on a 30,000-probe microarray—not exactly a phenotype in the classic sense, but as the direct result of gene transcription, production of RNA is a highly heritable trait. With expression measurements for each line of flies in the MA and G panels, the authors could estimate selection against mutations in an expression trait as the ratio of the variation among MA lines to the variation among G lines.

Rather than try to identify pleiotropic effects among every possible pair of loci in their data set, McGuigan et al. drew random sets of five expression traits, and estimated the mutational covariance among them (in the MA lines) and the standing genetic covariance among them (in the G lines) to calculate, finally, the selection coefficients acting on those pleiotropic effects. This wouldn’t produce an exhaustive estimate of pleiotropic selection, but it should give a good genome-wide sample.

The estimated selection coefficients for pleiotropic effects among random sets of five traits were almost an order of magnitude greater than the coefficients for individual traits. The whole distribution of selection coefficients for pleiotropic effects was shifted to higher values than the distribution for individual expression traits—17% of individual traits had selection coefficients greater than 0.02, while almost half (46%) of the pleiotropic effects did.

So it looks like this experimental system confirms a basic assumption of quantitative genetics, which is a nice result to see. Whether similar effects would be detectable in less well-controlled species than Drosophila remains to be seen, but the basic approach should be workable for organisms beyond fruit flies.

Reference

McGuigan K, JM Collet, SL Allen, SF Chenoweth, and MW Blows. 2014. Pleiotropic mutations are subject to strong stabilizing selection. Genetics. doi: 10.1534/genetics.114.165720.

RedditDiggMendeleyPocketShare and Enjoy

About Jeremy Yoder

Jeremy Yoder is a postdoctoral associate in the Department of Plant Biology at the University of Minnesota. He also blogs at Denim and Tweed and Nothing in Biology Makes Sense!, and tweets under the handle @jbyoder.
This entry was posted in adaptation, genomics, quantitative genetics and tagged , . Bookmark the permalink.