Adaptation to salinity affects species distributions, promotes speciation, and guides many evolutionary patterns in fishes. To uncover the basis of a complex trait like osmoregulation, genome-level analyses are sensible. We combine population genomic scans with genome expression profiling to discover candidate genes and pathways associated with divergence between osmotic environments. We compared transcriptome sequence divergence between multiple freshwater and saltwater populations of the rainwater killifish, Lucania parva. We also compared sequence divergence between L. parva and its sister species, Lucania goodei, a freshwater specialist. We found highly differentiated single nucleotide polymorphisms (SNPs) between freshwater and saltwater L. parva populations in cell junction and ion transport genes, including V-type H+ ATPase. Between species, we found divergence in reproduction and osmotic stress genes. Genes that were differentially expressed between species during osmotic acclimation included genes involved in ion transport and cell volume regulation. Gene sets that were divergent in coding sequence and divergent in expression did not overlap, although they did converge in function. Like many studies using genomic scans, our approach may miss some loci that contribute to adaptation but have complicated patterns of allelic variation. Our study suggests that gene expression and coding sequence may evolve independently as populations adapt to a complex physiological challenge.
Velvet_Reference_protein_matching.txt
Sequences of our reference based on the Lucania parva velvet assembly (kmer=25) with a total of 24,781 contigs, combining 1,657 contigs from the L. goodei assembly (all unique contigs over 200bp) with L. parva contigs that had blast matches to any of the following: 1) related teleost proteomes (Atlantic killifish: Fundulus heteroclitus, Japanese medaka: Oryzia latipes, three-spined stickleback: Gasterosteus aculeatus), 2) Fundulus ESTs on the microarray, or 3) target genes of interest (osmoregulation and vision related genes). Blast results from blastX searches against the uniprot proteomes were retained only if contigs had only 1-2 protein matches per species (blast score > 100) to avoid inclusion of contigs that were misassembled from multiple genes.
Velvet_Reference_protein_matching_rev.txt
Popoolation_Fst_all
Results from Fst calculation in Popoolation2 (Kofler et al. 2011). We calculated allele frequencies and Fst between all populations, using a minimum coverage of 8 reads per population (to exclude SNPs from transcripts expressed at low frequency), a pool size of 10, and a sliding window of 50bp. Columns show Fst for populations listed at top (for given 50bp SNP window), last column is average of all L. goodei and all L. parva comparisons. Contigs match velvet reference names. L. goodei populations: Upper Bridge (Wakulla River, Florida) and Everglades (26-Mile Bend, Florida). Freshwater L. parva populations: Pecos (Pecos River, Texas), Lower Bridge (Wakulla River, Florida), Delks (Delk's Bluff, Oklawaha River, Florida). Saltwater L. parva populations: Indian River (Atlantic Ocean, Florida Coast), St. Mark’s (Gulf of Mexico, Florida Coast), Bolivar Peninsula (Galveston Bay, Texas Coast).
Annotated_gene_universe_Lucania
List of annotations for all Lucania reference contigs using blastX and human proteome. Gene universe used in GO enrichment analyses.
Goodei_Parva_Outliers
List of annotations for outliers between species (L. goodei vs. L. parva) using blastX and human proteome. Target gene file used in GO enrichment analyses.
Goodei_outliers
List of annotations for outliers within L. goodei (Upper Bridge vs. Everglades) using blastX and human proteome. Target gene file used in GO enrichment analyses.
SWSW_outliers
List of annotations for outliers within L. parva between geographically distant saltwater populations (Indian River and Bolivar Peninsula) using blastX and human proteome. Target gene file used in GO enrichment analyses. These outliers were removed from all freshwater-saltwater drainage analyses.
Parva_outliers1Drain_SWrem
List of annotations for outliers between freshwater-saltwater drainages in L. parva using blastX and human proteome. Outliers in at least one drainage (and not present in saltwater-saltwater outliers). Target gene file used in GO enrichment analyses.
Parva_outliers2-3Drainages
List of annotations for outliers between freshwater-saltwater drainages in L. parva using blastX and human proteome. Outliers present in two to three drainages shown (and not present in saltwater-saltwater outliers). Target gene file used in GO enrichment analyses.
Infinium Chip Souce Sequences
Source sequences for Illumina Infinium Bead Chip custom designed for Lucania. 4,545 SNPs from 454 contigs (assembled using Newbler). Contains 1,679 candidate SNPs segregating between species, 1,497 SNPs segregating between Everglades and Upper Bridge L. goodei, 1,369 SNPs segregating between Florida Atlantic saltwater and Texas Gulf freshwater L. parva).
Infinuim_Chip_annotations
Annonations for Infinium contigs using BlastX against human proteome. Used in GO Enrichment analyses.
Infinium_Parva_Goodei_Fst
Fst calculations between 33 L. goodei and 39 L. parva genotyped on the Infinium Chip. Calculations done in hierfstat in R.
InfiniumFWSW_fst
Fst calculations between all freshwater and all saltwater L. parva (18 freshwater, 21 saltwater individuals: Florida Atlantic = 5 fresh, 6 salt; Florida Gulf = 8 fresh, 7 salt; Texas Gulf = 5 fresh, 8 salt) genotyped on the Infinium Chip. Calculations done in hierfstat in R.