Limits...
Outlier analyses to test for local adaptation to breeding grounds in a migratory arctic seabird

View Article: PubMed Central - PubMed

ABSTRACT

Investigating the extent (or the existence) of local adaptation is crucial to understanding how populations adapt. When experiments or fitness measurements are difficult or impossible to perform in natural populations, genomic techniques allow us to investigate local adaptation through the comparison of allele frequencies and outlier loci along environmental clines. The thick‐billed murre (Uria lomvia) is a highly philopatric colonial arctic seabird that occupies a significant environmental gradient, shows marked phenotypic differences among colonies, and has large effective population sizes. To test whether thick‐billed murres from five colonies along the eastern Canadian Arctic coast show genomic signatures of local adaptation to their breeding grounds, we analyzed geographic variation in genome‐wide markers mapped to a newly assembled thick‐billed murre reference genome. We used outlier analyses to detect loci putatively under selection, and clustering analyses to investigate patterns of differentiation based on 2220 genomewide single nucleotide polymorphisms (SNPs) and 137 outlier SNPs. We found no evidence of population structure among colonies using all loci but found population structure based on outliers only, where birds from the two northernmost colonies (Minarets and Prince Leopold) grouped with birds from the southernmost colony (Gannet), and birds from Coats and Akpatok were distinct from all other colonies. Although results from our analyses did not support local adaptation along the latitudinal cline of breeding colonies, outlier loci grouped birds from different colonies according to their non‐breeding distributions, suggesting that outliers may be informative about adaptation and/or demographic connectivity associated with their migration patterns or nonbreeding grounds.

No MeSH data available.


Plots of clustering analyses. (a) DAPC results for all loci and (b) only outliers; plots (c) and (d) are based on the ‘only outliers’ dataset; (c) plot showing Evanno's ΔK with two peaks corresponding to K = 3 and 5; (d) barplots of STRUCTURE results for K = 3, which received highest support from Evanno's ΔK
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5383466&req=5

ece32819-fig-0002: Plots of clustering analyses. (a) DAPC results for all loci and (b) only outliers; plots (c) and (d) are based on the ‘only outliers’ dataset; (c) plot showing Evanno's ΔK with two peaks corresponding to K = 3 and 5; (d) barplots of STRUCTURE results for K = 3, which received highest support from Evanno's ΔK

Mentions: Results from structure and DAPC were highly concordant when applied to thick‐billed murres only. When all SNPs were included, structure failed to detect any population structure (data not shown). Similarly, the DAPC showed a high degree of overlap among colonies (Figure 2a). However, results from analyses that included only SNPs potentially under positive selection were very different. Both structure and DAPC grouped the southernmost colony, Gannet, with the two northernmost colonies, Minarets and Prince Leopold, whereas individuals from Coats and Akpatok clustered separately from each other and from any other colonies (Figure 2b,d). The graphical results were supported by Evanno's ΔK, which indicated highest support for three distinct populations (Figure 2c). ΔK also showed a peak at K = 5, indicating that individuals are more similar within colonies than among colonies and suggesting hierarchical population structure (Figure 2b,d). Interestingly, two individuals from Akpatok were assigned to the Minarets population in structure and overlapped with individuals from Minarets in the DAPC as well. Because these two individuals could represent migrants from Minarets into Akpatok, we excluded them from the dataset and repeated the DAPC. When those two samples were removed from the analyses, Akpatok separated more strongly from the other colonies, making the DAPC results match the structure results (Fig. S4). lositan identified a total of 252 outlier SNPs from the randomized dataset, but both structure and the DAPC failed to detect population structure in this dataset, thus indicating that the apparent distinctiveness of the colonies is not an artifact of the analyses.


Outlier analyses to test for local adaptation to breeding grounds in a migratory arctic seabird
Plots of clustering analyses. (a) DAPC results for all loci and (b) only outliers; plots (c) and (d) are based on the ‘only outliers’ dataset; (c) plot showing Evanno's ΔK with two peaks corresponding to K = 3 and 5; (d) barplots of STRUCTURE results for K = 3, which received highest support from Evanno's ΔK
© Copyright Policy - creativeCommonsBy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5383466&req=5

ece32819-fig-0002: Plots of clustering analyses. (a) DAPC results for all loci and (b) only outliers; plots (c) and (d) are based on the ‘only outliers’ dataset; (c) plot showing Evanno's ΔK with two peaks corresponding to K = 3 and 5; (d) barplots of STRUCTURE results for K = 3, which received highest support from Evanno's ΔK
Mentions: Results from structure and DAPC were highly concordant when applied to thick‐billed murres only. When all SNPs were included, structure failed to detect any population structure (data not shown). Similarly, the DAPC showed a high degree of overlap among colonies (Figure 2a). However, results from analyses that included only SNPs potentially under positive selection were very different. Both structure and DAPC grouped the southernmost colony, Gannet, with the two northernmost colonies, Minarets and Prince Leopold, whereas individuals from Coats and Akpatok clustered separately from each other and from any other colonies (Figure 2b,d). The graphical results were supported by Evanno's ΔK, which indicated highest support for three distinct populations (Figure 2c). ΔK also showed a peak at K = 5, indicating that individuals are more similar within colonies than among colonies and suggesting hierarchical population structure (Figure 2b,d). Interestingly, two individuals from Akpatok were assigned to the Minarets population in structure and overlapped with individuals from Minarets in the DAPC as well. Because these two individuals could represent migrants from Minarets into Akpatok, we excluded them from the dataset and repeated the DAPC. When those two samples were removed from the analyses, Akpatok separated more strongly from the other colonies, making the DAPC results match the structure results (Fig. S4). lositan identified a total of 252 outlier SNPs from the randomized dataset, but both structure and the DAPC failed to detect population structure in this dataset, thus indicating that the apparent distinctiveness of the colonies is not an artifact of the analyses.

View Article: PubMed Central - PubMed

ABSTRACT

Investigating the extent (or the existence) of local adaptation is crucial to understanding how populations adapt. When experiments or fitness measurements are difficult or impossible to perform in natural populations, genomic techniques allow us to investigate local adaptation through the comparison of allele frequencies and outlier loci along environmental clines. The thick‐billed murre (Uria lomvia) is a highly philopatric colonial arctic seabird that occupies a significant environmental gradient, shows marked phenotypic differences among colonies, and has large effective population sizes. To test whether thick‐billed murres from five colonies along the eastern Canadian Arctic coast show genomic signatures of local adaptation to their breeding grounds, we analyzed geographic variation in genome‐wide markers mapped to a newly assembled thick‐billed murre reference genome. We used outlier analyses to detect loci putatively under selection, and clustering analyses to investigate patterns of differentiation based on 2220 genomewide single nucleotide polymorphisms (SNPs) and 137 outlier SNPs. We found no evidence of population structure among colonies using all loci but found population structure based on outliers only, where birds from the two northernmost colonies (Minarets and Prince Leopold) grouped with birds from the southernmost colony (Gannet), and birds from Coats and Akpatok were distinct from all other colonies. Although results from our analyses did not support local adaptation along the latitudinal cline of breeding colonies, outlier loci grouped birds from different colonies according to their non‐breeding distributions, suggesting that outliers may be informative about adaptation and/or demographic connectivity associated with their migration patterns or nonbreeding grounds.

No MeSH data available.