It has been found that sex-biased genes, those more expressed in one sex than in the other, are not randomly distributed on the chromosomes in Drosophila [1–3]. Male-biased genes are generally under-represented on the X chromosome, except the very young genes, whereas female-biased genes are enriched on the X [1, 3]. In addition, there is an excess of gene movement from the X chromosome to autosomal locations, with new retrogenes acquiring testis-biased expression pattern . Those two related phenomena have been broadly observed in the Drosophila genus [4–9], in mosquitos [10–12], and mammals [13, 14]. The X chromosomes from all Drosophila species analyzed, including Neo-X chromosomes, were found to be under-represented with male-biased genes [4, 5]. Further, the excess movement of retrogenes and DNA-based duplications off the X chromosome was observed in 12 Drosophila species whose genomes were sequenced [8, 9]. In Drosophila, gene movement off the X chromosome was suggested to be a mechanism by which the autosomes become enriched with male-biased genes .
These observations raise interesting questions about the processes shaping sex chromosome evolution, particularly the relationship between male-biased gene expression and the under-representation of this class of genes on the X chromosome. Over the past decade, four hypotheses, including sexual antagonism, meiotic sex chromosome inactivation, dosage compensation, and meiotic drive, have been proposed to interpret the paucity of male-biased X-linked genes [2, 15–23]. The first hypothesis, sexual antagonism, assumes that sexually antagonistic forces drive male-biased expression. In such case, the X chromosome, which is present in a single copy in males compared to two copies in females, would have less opportunity to accumulate male-biased genes [15, 16, 21]. More specifically, sexually antagonistic dominant mutations with male-beneficial and female-detrimental effects have a higher probability of fixation on the autosomes [15, 16]. However, a recent study has shown that sexual antagonistic genes tend to be preferentially located on the X chromosome . This result suggests that sex-biased genes are not currently under sexual antagonistic selection but rather represent the partial or total resolution of the phenomenon . The second hypothesis, dosage compensation, predicts that the hypertranscription of the X chromosome in Drosophila could further limit the up-regulation of genes and therefore prevent the origination of male-biased genes on the X chromosome [18, 19]. The third hypothesis proposes that meiotic sex chromosome inactivation (MSCI) could favour the accumulation of testis-biased genes in the autosomes [2, 20]. Different from X-linked genes, autosomal genes are free from the inactivation process and therefore have an increased probability of being expressed in males [2, 20]. In the fourth hypothesis, meiotic drive alleles located on X chromosome and expressed during spermatogenesis could favour the evolution of autosomal male-biased genes as their potential suppressors [22, 23].
Empirical evidence exists in support of most of these hypotheses suggesting that all of them may have played a role in chromosomal distribution of male-biased genes [1, 18, 19, 24, 25]. Evidence supporting the sexual antagonism hypothesis comes from the observation of the paucity of X-linked male-biased genes expressed in somatic tissues which do not undergo X chromosome inactivation , whereas evidence supporting the dosage compensation hypothesis comes from studies showing that: (1) male-biased genes are less likely to be bound by the MSL complex ; and (2) highly expressed male-biased genes are more rarely found on the X chromosome .
MSCI has been shown to occur in a wide range of taxa: mammals, nematodes, chicken, and Drosophila [20, 24–29]. Although there is unequivocal evidence for MSCI in mammals, until recently the only indirect evidence for MCSI in Drosophila was from the pioneering work of Lifschytz and Lindsley . There are now two major lines of supporting evidence for MSCI in Drosophila [24, 25, 29, 30]. First, insertion into the X chromosome of genes carrying a testis-specific promoter had reduced expression compared to the same insertions into autosomes , a result consistent with the MSCI model. These results were further confirmed by a more exhaustive study of insertions across different regions of the entire X chromosome . Second, a global analysis of gene expression between testis samples enriched with mitotic and meiotic cells showed a significant down-regulation of the X chromosome in agreement with MSCI . Yet, a recent study argues that this X chromosome-specific down-regulation starts in earlier stages of the mitotic male germline .
Nonetheless, MSCI was demonstrated to be one of the driving forces for the genomic relocation of testis-biased genes . First, the under-representation of testis-biased genes was found for genes over-expressed in meiosis, but not in mitosis . Second, parent-retrogene pairs moving out of the X chromosome have higher complementary expression in meiosis, that is parental gene down-regulation and retrogene up-regulation, than those pairs moving between autosomes . Those observations directly link the testis-biased X chromosome deficiency to a meiotic event as expected by MSCI in males.
However, a recent study using an alternative approach to assess MSCI in Drosophila claimed that there was no sign of X inactivation during male meiosis . Different larval development stages were used to obtain testis with differing amounts of spermatogenic meiotic cells . No differential expression between autosomes and X chromosomes was detected during larval development and therefore the global X inactivation in male germline was ruled out as a possible process . The same study , using the public Drosophila expression dataset  analyzed the chromosomal distribution of tissue-specific genes and found that several non-sex-related tissues, besides the testis as previously thought [1, 3–5], had paucity of X-linked genes. Taken together, the authors suggested that there was no evidence for MSCI and therefore could not be a driving force behind the chromosomal distribution of male-biased genes .
To better understand the difference between these analyses and previous conclusions, we re-analyzed the data of this recently published study . First, we found that the larval testis data generated by Mikhaylova and Nurminsky  have low within-replicate correlations, which should make the detection of differential chromosomal expression practically impossible. Second, we also found that the tissue-specific gene datasets used by Mikhaylova and Nurminsky  were actually enriched with testis-biased genes. Using a non-enriched dataset after filtering out the testis-specific genes, we found that non-sex-biased tissue-specific genes were not under-represented on the X chromosome. In the sections below, we report the details of our re-analyses.