Many salmonid species exist in highly structured and isolated populations, and are susceptible to habitat fragmentation and disturbances. Gila Trout (Oncorhynchus gilae) is a threatened species found in the Southwestern United States, and is managed to preserve relict populations (i.e., lineages) distributed across a fragmented landscape. We evaluated genomic variation within and among remaining lineages of Gila Trout using RADseq to assess how drift and selection have structured populations using neutral and outlier loci. We also examined whether a signature of hybridization was evident in relict populations. Gila Trout lineages were significantly differentiated from one another and were characterized by low effective population sizes. However, most lineages maintained genomic diversity and are potentially adapted to local conditions. Hybridization with non-native Rainbow Trout (O. mykiss) was not detected in any lineage. All lineages have experienced population bottlenecks associated with mortality from drought and severe wildfires. Current management strategies should be reevaluated and adapted to better account for long-term effects of climate change. Specifically, we suggest reconnecting some populations via dendritic stream networks to facilitate natural dispersal in a metapopulation context. This would allow natural genetic mixing on the landscape and potentially increase adaptive potential. Furthermore, genetic rescue should be implemented to preserve integrity of the unique Spruce Creek lineage that is currently compromised by extremely low diversity.

Gila Trout fin clips were obtained through the Museum of Southwestern Biology (MSB) at the University of New Mexico (https://msb.unm.edu). Gila Trout samples (n=154) representing all five wild-origin, relict lineages were assessed, and total genomic DNA was isolated from fin clip samples. Genomic DNA was submitted to SNPsaurus (http://snpsaurus.com) where genomic libraries were prepared and a reduced-representation nextRAD sequencing method was applied to amplify genomic loci and build RAD tag libraries (150 bp reads and average 20X depth). Raw reads were trimmed and mapped to an annotated Rainbow Trout genome (https://www.ncbi.nlm.nih.gov/genome/?term=rainbow+trout), and genotype genetic variants. Resulting SNP loci were filtered for sequencing errors, genotyping errors, linkage disequilibrium, and paralogs. Only biallelic SNPs were considered. Outlier loci were removed from the neutral dataset and retained for the outlier dataset. Outliers were assessed with two methods (Bayescan and PCAdapt) and estimated outliers were annotated. Only outliers annotated to Rainbow Trout genome were considered. Genotypes were assessed with VCFtools.

Genotypes are coded as 0, 1, or 2 depending on the number of alternate alleles each individual is contained (0 = homozygous reference allele, 1 = heterozygotes, 2 = homozygous alternate allele). CHROM ID's refer to RefSeq sequence ID's of Oncorhynchus gilae (Rainbow Trout) found at https://www.ncbi.nlm.nih.gov/assembly/GCF_002163495.1