[This author post is by Alex Kalinka and Pavel Tomancak on their paper, An excess of gene expression divergence on the X chromosome in Drosophila embryos: implications for the faster-X hypothesis, posted to the arXiv here.]

We have been working towards publishing our study of gene expression evolution on the X chromosome in Drosophila embryos since the beginning of March this year. Recently, Casey Bergman suggested that we upload our manuscript to the arXiv, and after we did so, we were kindly invited by Graham Coop to write a guest post about our work for Haldane's Sieve.

It makes sense to post here since the roots of our study go back to Haldane in 1924 [1]; he recognised that the unusual inheritance pattern of the X chromosome, in which a single copy is present in the heterogametic sex versus two copies in the homogametic sex, could in turn lead to unusual evolutionary patterns on the X relative to the autosomes. If, for example, a beneficial mutation is recessive, then it would be more exposed to natural selection in the heterogametic sex where, relative to an equivalent autosomal allele, it would spend less time being masked by the dominant, less beneficial allele [1]. The prediction that adaptive evolution might proceed more quickly on the X than on the autosomes has been dubbed the faster-X hypothesis. However, the X chromosome might also be expected to evolve more rapidly for non-adaptive reasons. In each mating pair there will be 3 copies of the X chromosome versus 4 copies of each autosome, which might in turn lead to a lower chromosomal effective population size for the X thereby increasing the strength of random genetic drift.

While some studies have reported evidence for a faster-X effect for adaptive protein evolution in Drosophila, other studies have reported that there is no difference between the X and the autosomes, and to date the evidence is somewhat inconclusive. As we focused our study on gene expression, we had an opportunity to relax the implicit assumption that the majority of adaptive evolution occurs in coding regions. To help disentangle adaptive and non-adaptive evolutionary signatures in our data, we used both between-species measures of gene expression divergence and within-species measures of gene expression variation using inbred strains of D. melanogaster generated by the Drosophila Genetic Reference Panel (DGRP).

We found an excess of gene expression divergence on the X chromosome between six Drosophila species (a mean increase of ~20%). In contrast, we found that the X exhibits a significantly lower level of gene expression variation between inbred strains of D. melanogaster (a mean decrease of ~10%). Taken together, these results suggest that the divergence that we find between species is not driven by a relaxation of selective constraint on the X chromosome. To further explore whether such a signature could be driven by the hemizygosity of the X, we analysed gene expression in mutation accumulation lines of D. melanogaster. If the single copy of the X in males is driving the excess of expression divergence that we found on the X, then we would expect to find an excess of expression variation between lines that have independently accumulated mutations. In fact, we found the opposite was the case – the X chromosome displayed a significantly lower rate of mutation accumulation than the autosomes suggesting that the hemizygosity of the X alone is not sufficient to drive a higher rate of fixation of gene expression mutations.

Overall, we argue that the excess divergence we find on the X is best understood within the framework of the faster-X hypothesis. In support of this interpretation, we find that there is an excess of gene expression divergence on Muller's D element along the branch leading to the obscura sub-group; Muller's D element segregates as a neo-X chromosome in the obscura sub-group, and hence provides a powerful, independent test of faster evolution of the X chromosome.

Several questions remain, however, and we hope that our findings will help to stimulate further research into the details underpinning the differences we find on the X. In particular, work needs to be done to discover the genetics underlying divergence on the X, such as the relative importance of cis versus trans-acting factors, and, crucially, we need to develop a better understanding of how variation in gene expression impacts organismal fitness. Research into the latter question is essential if we are to bridge the conspicuous gaps between sequence variation, gene expression variation, and organismal fitness.

Dave Gerrard initially found elevated gene expression divergence on the X in the course of analysing data for our developmental hourglass paper, and spoke about his findings at the 43rd population genetics conference; that was more than two years ago. Since then we collected new data, and it took a while to put the paper together although it is still not certain that it will be published in a traditional journal. The arXiv is a great way to let the scientific community know about your results before the academic process runs its course. We only regret we didn't make use of this excellent outlet back in March.

[1] Haldane JBS (1924) A mathematical theory of natural and artificial selection. Part I. Trans Camb Phil Soc 23: 19-41.