Phenotypic plasticity is expected to facilitate the persistence of natural populations as global change progresses. The attributes of fluctuating environments that favor the evolution of plasticity have received extensive theoretical investigation, yet empirical validation of these findings is still in its infancy. Here, we combine high-resolution environmental data with a laboratory-based experiment to explore the influence of habitat pH fluctuation dynamics on the plasticity of gene expression in two populations of the Mediterranean mussel, Mytilus galloprovincialis. We linked differences in the magnitude and predictability of pH fluctuations in two habitats to population-specific gene expression profiles in ambient and stressful pH treatments. The results presented demonstrate population-based differentiation in gene expression plasticity, whereby mussels native to a habitat exhibiting a large magnitude of pH fluctuations with low predictability display reduced phenotypic plasticity between experimentally imposed pH treatments. This work validates recent theoretical findings on evolution in fluctuating environments using an ecologically important marine bivalve, and suggests that populations inhabiting regions exposed to unpredictably fluctuating selection pressures may exhibit reduced plasticity as global change progresses.

For a full description of the methods, please see the associated manuscript for this Dryad submission.

Breifly, one year of high-frequency pH and temperature monitoring was conducted at a coastal and lagoon habitat in the Northwest Mediterranean Sea (the Bay of Villefranche: 43.682^oN, 7.319^o E and Thau Lagoon: 43.4158^oN, 3.6888^oE). Resident mussels at each site were collected and reared in labortory common garden conditions for six weeks, after which a 5 day acclimation to two pH treatments (benign pH - 8.1; stressful pH - 7.75) was conducted. At the end of the acclimation peirod, total RNA was extracted from gill tissue samples,  and gene expression profiling was carried out using 3’ mRNA sequencing. All resulting raw and processed data files are included here. 

TranscriptCounts.csv – Transcript count data for all coastal and lagoon mussels reared in either benign or stressful pH conditions. Each row corresponds to a gene, while each column corresponds to the observed transcript count for each individual at that gene. Transcript counts were computed from the sequencing data using custom Perl script written by Misha Matz (available at https://github.com/z0on/tag-based_RNAseq).

CountsID.csv – Identifying information for the individuals in the transcript counts matrix. Specifically, each row contains pertinent information for corresponding column of the transcript counts file, including source population and treatment information.

SupplementaryFile1.csv – File containing gene expression data for the coastal population’s response to stressful pH conditions. Columns correspond to: reference contig ID (as identified in reference transcriptome provided by Moreira et al. 2015), transcript ID (corresponding to row number of counts matrix), log2-fold change, and raw p-value for each gene. Log2-fold-change values were computed using the DeSEQ2 software developed by Love et al. (2015) (doi:10.1186/s13059-014-0550-8).

SupplementaryFile2.csv – File containing gene expression data for the lagoon population’s response to stressful pH conditions. Columns correspond to: reference contig ID (as identified in reference transcriptome provided by Moreira et al. 2015), transcript ID (corresponding to row number of counts matrix), log2-fold change, and raw p-value for each gene. Log2-fold-change values were computed using the DeSEQ2 software developed by Love et al. (2015) (doi:10.1186/s13059-014-0550-8).

SupplementaryFile3.csv – File containing gene expression data for population differentiation in the benign pH treatment. Columns correspond to: reference contig ID (as identified in reference transcriptome provided by Moreira et al. 2015), transcript ID (corresponding to row number of counts matrix), log2-fold change, and raw p-value for each gene. Log2-fold-change values were computed using the DeSEQ2 software developed by Love et al. (2015) (doi:10.1186/s13059-014-0550-8).

SupplementaryFile4.csv – File containing gene expression data for population differentiation in the stressful pH treatment. Columns correspond to: reference contig ID (as identified in reference transcriptome provided by Moreira et al. 2015), transcript ID (corresponding to row number of counts matrix), log2-fold change, and raw p-value for each gene. Log2-fold-change values were computed using the DeSEQ2 software developed by Love et al. (2015) (doi:10.1186/s13059-014-0550-8).

SupplementaryFile5.csv – File containing temperature and pH data as measured by SeaFET pH sensor at hourly intervals from the coastal habitat.

SupplementaryFile6.csv – File containing raw SeaFET pH sensor readings, as well as sensor-collected pH and temperature data at 20 minute intervals from the lagoon habitat.