An evaluation of the role of adaptation in salmon evolution using genome based approaches

Studying the results of selection may provide insights into the extent of adaptation, processes affecting population divergence, and gene diversity. Here, the role of adaptation in salmon evolution was evaluated at different taxonomic levels using genome based approaches. The first part of this thesis was aimed at developing a bioinformatic methodology to detect genes under selection on a large scale in non-model species. In such species, coding sequences can be incomplete because of limited genomic resources. However, these sequences are information rich, and can be used to estimate neutral versus non neutral divergence across species. Incomplete DNA sequences can complicate estimates of non-neutral divergence based on comparisons between synonymous (dS) and non-synonymous (dN) nucleotide substitutions, commonly used to study selection between species. The first chapter describes a series of steps that can be used to examine positive selection on a large scale between non model species using partial sequences. The methodology is described for six species of salmonids, where approaches are complicated by the fact that a whole duplication event occurred in the lineage leading to these species. Therefore, challenges associated with duplicated genomes, specifically the separation of orthologs from paralogs, were also addressed. We found that multi-way BLAST optimized the number of alignments between partial coding sequences. We recommend that reading frames should be manually detected after alignment with sequences in Genbank using the BLASTX program. Finally, phylogenetic approaches were determined to be suitable to separate orthologs from paralogs in duplicated genomes. The second part of the thesis was aimed at conducting a genome-wide assessment of the role of adaptive evolution of Chinook salmon in the Columbia River in the Pacific Northwest of the USA. The first step involved the construction of a dense linkage map for Chinook salmon, thus providing the necessary resources for a genome-wide analysis in wild populations (Chapter 2). We mapped 7146 RAD loci on the 34 chromosomes of Chinook salmon, spanning 4163cM. All the chromosome arms were identified through centromere mapping. Placement of 799 duplicated loci revealed that they were preferentially distributed on distal regions of eight pairs of chromosome arms. This result suggests that homeologs diverged at different rates following whole genome duplication. Our results supported near complete interference during recombination for Chinook salmon, and confirmed previously identified homologies between Chinook salmon and rainbow trout. In the third chapter, we aimed to determine the role of adaptive divergence in the evolution of Chinook salmon in the Columbia River Basin. A population survey of divergence was conducted using 14105 RAD markers in eleven populations in the Columbia River Basin, representative of the three main lineages identified in previous studies. Our results supported the hypothesis of colonization of the Columbia River Basin from two main refugia following the last glaciation event. We identified 301 outlier loci that did not conform to neutral evolution, consistent with adaptive divergence. Of these, 148 and 153 were associated with the pre- and post- glaciation divergence respectively. Using the linkage map created in the second chapter, we identified chromosomal regions of high divergence, most of which were located in distal regions from the centromere. Although some regions of elevated divergence were observed in common between lineages, many appeared to be specific to pre- or post-glaciation divergence. Finally we investigated whether we could find molecular evidence supporting observations of parallel evolution in a phenotypic trait across populations, adult return timing. Random forest analyses, a regression-based approach, detected some loci that predicted run timing, specifically Spring and Fall return timing, two of which mapped to the same position on the linkage map. In this chapter, beyond improving our understanding of Chinook salmon evolution, we have demonstrated the usefulness of dense linkage maps in identifying regions of the genome that may have been involved in adaptive evolution. The research presented in this thesis will facilitate the study of adaptive divergence between non-model species. Novel and extensive genomic resources for Chinook salmon have also been developed. These resources have provided insights into chromosome evolution following whole genome duplication, and have greatly contributed to the understanding of adaptive evolution of populations of Chinook salmon in the Columbia River Basin.