Back

Sex-biased expression in whole bodies, tissues and cell-types: patterns across and within levels

Liu, M. J.; Panyam, S.; Agrawal, A. F.

2026-07-20 evolutionary biology
10.64898/2026.07.14.738470 bioRxiv
Show abstract

Whole body transcriptomes have been widely used to study the evolution of sex-biased genes, yet it remains unclear what whole body sex bias represents biologically. Using the FlyAtlas2 and Fly Cell Atlas datasets from Drosophila melanogaster, we show that whole body sex bias emerges from both sex-biased regulation within smaller biological units and how genes are expressed these across units, the latter indicating the contribution of compositional differences to whole body sex bias. Despite these compositional effects, the direction of whole body sex bias is generally consistent with that observed across most tissues and cell-types because sex bias tends to be positively correlated among tissues and among cell-types. Genes which display sex-biased expression across multiple tissues/cell-types typically exhibit the greatest magnitude of sex bias in reproductive tissues/cell-types, consistent with widespread pleiotropic spillover from reproductive components. Among genes lacking gonadal sex bias, the magnitude of sex-biased expression is often greatest in tissues where genes are less highly expressed, further suggesting the role of pleiotropy. In revisiting the evolutionary patterns with respect to tissue level sex bias, we find that elevated rates of adaptive evolution among gonad-expressed sex-biased genes are better explained by their greater expression localization in the reproductive tissues, rather than by their sex bias itself. However, among genes not expressed in the gonads, strong male-biased expression in non-gonadal tissues remains independently associated with increased adaptive evolution. Together, our results show that whole body sex bias is a meaningful summary of broad-scale sex-biased expression though may obscure finer-scale evolutionary signals.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.