PgmNr P322: Genome-wide selection component analysis in a wild pedigreed population of the Florida Scrub-Jay.

Authors:
Andrew G. Clark 1 ; Elissa Cosgrove 1 ; Huijie Feng 1 ; Ishaan Jhaveri 1 ; Ashish Akshat 1 ; Reed Bowman 2 ; John W. Fitzpatrick 3 ; Nancy Chen 3,4


Institutes
1) Cornell Univ, Ithaca, NY; 2) Archbold Biological Station, Venus, FL; 3) Cornell Lab of Ornithology, Ithaca, NY; 4) University of California, Davis, CA.


Abstract:

Analysis of contemporary evolution is the only way to directly test many fundamental questions in evolutionary biology, but we usually lack the combined phenotypic and genomic data over time in natural populations required for such studies. In addition, most current approaches for inferring natural selection do not have the opportunity to incorporate the additional power gained from access to a full pedigree. Here, we study short-term selection using a 25-year genomic, phenotypic, and pedigree dataset in the Florida Scrub-Jay (Aphelocoma coerulescens), an iconic species on the U. S. Endangered Species List that has drastically decreased in number during the past half-century. A population of Florida Scrub-Jays at Archbold Biological Station has been studied intensively since 1969, providing a 12-generation pedigree among the most accurate and extensive for any wild vertebrate species. For all locally recruited birds in the population, we have full records of individual lifespans, annual fecundity, and lifetime fitness; similar records exist for immigrant breeders except for exact lifespan. We sequenced and assembled the Florida Scrub-Jay genome and used custom Illumina Beadchips to genotype every individual in our study population over the past two decades (3,838 individuals total) at 15,416 genome-wide SNPs. We used gene dropping to explicitly sample gametes in each generation on the known pedigree and asked whether the observed allele frequency dynamics of each SNP were consistent with a pure drift process. Numerous SNPs departed significantly from the null model and showed frequency dynamics consistent with perturbation by selection. We then tested for selection acting on specific life-cycle stages by tailoring hierarchical selection component analysis to take full advantage of exhaustive population sampling. We identified a number of loci that clearly exhibited male gametic selection, sexual selection, and viability selection. By combining sensitive pedigree-based inferences of net selection with fine-scale dissection of selection components, this study provides a detailed assessment of the role of selection in perturbing allele frequency dynamics in a rapidly declining population. Results suggest a role of selection in maintaining variation even in the face of population decline, and may help guide conservation efforts.