Back

Evolutionary constraint predicts the repeatability of transcriptional divergence during repeated adaptation

Tobias, Z. J. C.; Rennison, D.

2026-10-02 evolutionary biology
10.64898/2026.09.28.755181 bioRxiv
Show abstract

The repeated evolution of similar phenotypes in response to shared ecological pressures provides striking examples of natural selection and raises broader questions about the predictability of evolution. Evolutionary constraint may shape this predictability, while the apparent extent of parallelism may also depend on the biological scale at which it is measured. When are adaptive solutions channeled along the same route? We addressed this question using three sympatric benthic-limnetic ecotype pairs of threespine stickleback (Gasterosteus aculeatus), characterizing gene expression in gill and liver from wild-caught fish. We tested whether proxies of pleiotropic and coding-sequence constraint predicted the repeatability of transcriptional divergence across lakes. At the gene level, differential expression was strongly biased toward parallel rather than antiparallel responses. At the functional level, numerous pathways showed parallel enrichment, but the conservation of their underlying member-gene responses varied substantially, indicating that similar functional outcomes can arise with differing degrees of gene-level concordance. Most importantly, several proxies of evolutionary constraint predicted transcriptional repeatability, particularly in gill: broadly expressed, highly connected, and more slowly evolving genes showed more stable expression responses across lakes, with similar relationships evident for repeated use of genes within functionally parallel pathways. Together, these results show that both gene-level constraint and biological scale contribute to variation in molecular repeatability and suggest that the predictability of repeated adaptation may reflect an interaction between genetic architecture, population-genetic context, and the level of biological organization at which parallelism emerges.

Matching journals

The top 5 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.