The molecular basis of convergent phenotypes is often unknown. However, convergence at a genomic level is predicted when there are large population sizes, gene flow among diverging lineages or strong genetic constraints. We used whole-genome resequencing to investigate genomic convergence in fishes ( Poecilia spp.) that have repeatedly colonized hydrogen sulfide (H 2 S)-rich environments in Mexico. We identified genomic similarities in both single nucleotide polymorphisms (SNPs) and structural variants (SVs) among independently derived sulfide spring populations, with approximately 1.2% of the genome being shared among sulfidic ecotypes. We compared these convergent genomic regions to candidate genes for H 2 S adaptation identified from transcriptomic analyses and found that a significant proportion of these candidate genes (8%) were also in regions where sulfidic individuals had similar SNPs, while only 1.7% were in regions where sulfidic individuals had similar SVs. Those candidate genes included genes involved in sulfide detoxification, the electron transport chain (the main toxicity target of H 2 S) and other processes putatively important for adaptation to sulfidic environments. Regional genomic similarity across independent populations exposed to the same source of selection is consistent with selection on standing variation or introgression of adaptive alleles across divergent lineages. However, combined with previous analyses, our data also support that adaptive changes in mitochondrially encoded subunits arose independently via selection on de novo mutations. Pressing questions remain on what conditions ultimately facilitate the independent rise of adaptive alleles at the same loci in separate populations, and thus, the degree to which evolution is repeatable or predictable. This article is part of the theme issue ‘Convergent evolution in the genomics era: new insights and directions'.

Rapid glacier recession is altering the physical conditions of headwater streams. Stream temperatures are predicted to rise and become increasingly variable, putting entire meltwater‐associated biological communities at risk of extinction. Thus, there is a pressing need to understand how thermal stress affects mountain stream insects, particularly where glaciers are likely to vanish on contemporary timescales. In this study, we measured the critical thermal maximum (CTMAX) of stonefly nymphs representing multiple species and a range of thermal regimes in the high Rocky Mountains, USA. We then collected RNA‐sequencing data to assess how organismal thermal stress translated to the cellular level. Our focal species included the meltwater stonefly, Lednia tumana, which was recently listed under the U.S. Endangered Species Act due to climate‐induced habitat loss. For all study species, critical thermal maxima (CTMAX > 20°C) far exceeded the stream temperatures mountain stoneflies experience (<10°C). Moreover, while evidence for a cellular stress response was present, we also observed constitutive expression of genes encoding proteins known to underlie thermal stress (i.e., heat shock proteins) even at low temperatures that reflected natural conditions. We show that high‐elevation aquatic insects may not be physiologically threatened by short‐term exposure to warm temperatures and that longer‐term physiological responses or biotic factors (e.g., competition) may better explain their extreme distributions.

microRNAs (miRNAs) are post‐transcriptional regulators of gene expression and can play an important role in modulating organismal development and physiology in response to environmental stress. However, the role of miRNAs in mediating adaptation to diverse environments in natural study systems remains largely unexplored. Here, we characterized miRNAs and their expression in Poecilia mexicana, a species of small fish that inhabits both normal streams and extreme environments in the form of springs rich in toxic hydrogen sulphide (H2S). We found that P. mexicana has a similar number of miRNA genes as other teleosts. In addition, we identified a large population of mature miRNAs that were differentially expressed between locally adapted populations in contrasting habitats, indicating that miRNAs may contribute to P. mexicana adaptation to sulphidic environments. In silico identification of differentially expressed miRNA‐mRNA pairs revealed, in the sulphidic environment, the downregulation of miRNAs predicted to target mRNAs involved in sulphide detoxification and cellular homeostasis, which are pathways essential for life in H2S‐rich springs. In addition, we found that predicted targets of upregulated miRNAs act in the mitochondria (16.6% of predicted annotated targets), which is the main site of H2S toxicity and detoxification, possibly modulating mitochondrial function. Together, the differential regulation of miRNAs between these natural populations suggests that miRNAs may be involved in H2S adaptation by promoting functions needed for survival and reducing functions affected by H2S. This study lays the groundwork for further research to directly demonstrate the role of miRNAs in adaptation to H2S. Overall, this study provides a critical stepping‐stone towards a comprehensive understanding of the regulatory mechanisms underlying the adaptive variation in gene expression in a natural system.

7905 24 25 Running head: Thermal tolerance of mountain stoneflies 26 27 Abstract: 31 Rapid glacier recession is altering the physical conditions of headwater streams. Stream 32temperatures are predicted to rise and become increasingly, potentially putting entire meltwater-33

Eye regression occurs across cave-dwelling populations of many species and is often coupled with a decrease or loss in eye function. Teleost fishes are among the few vertebrates to undergo widespread colonization of caves and often exhibit eye regression with blindness. Cave populations of the poeciliid fish Poecilia mexicana (cave molly) exhibit reduced—albeit functional—eyes, offering the opportunity to investigate partial eye regression. We sequenced eye transcriptomes of cave and surface populations of P. mexicana to identify differentially expressed genes that potentially underlie eye regression in cave mollies. We identified 28 significantly differentially expressed genes, 20 of which were directly related to light sensitivity, eye structure and visual signaling. Twenty-six of these genes were downregulated in cave compared to surface populations. Functional enrichment analysis revealed eye-related gene ontologies that were under-represented in cave mollies. In addition, a set of co-expressed genes related to vision and circadian rhythm was correlated with habitat type (cave versus surface). Our study suggests that differential gene expression plays a key role in the beginning evolutionary stages of eye regression in P. mexicana , shedding further light on regressive evolution in cavefish.