W588 Search for Genomic Footprints of Divergent Selection From Whole Genome Sequencing of Four Experimental Lines of Broiler Chickens

Date: Sunday, January 15, 2012
Time: 8:30 AM
Room: Sunset
Behnam Abasht , Department of Animal and Food Sciences, University of Delaware, Newark, DE
Chuming Chen , Center for Bioinformatics and Computational Biology, University of Delaware, Newark, DE
Hongzhan Huang , Center for Bioinformatics and Computational Biology, University of Delaware, Newark, DE
Cathy H. Wu , Center for Bioinformatics and Computational Biology, University of Delaware, Newark, DE
Michel J. Duclos , Station de Recherches Avicoles, INRA, Nouzilly, France
Elisabeth LeBihan-Duval , Station de Recherches Avicoles, INRA, Nouzilly, France
Jean Simon , Station de Recherches Avicoles, INRA, Nouzilly, France
Larry A. Cogburn , Center for Bioinformatics and Computational Biology, University of Delaware, Newark, DE
We conducted whole genome sequencing of four experimental lines, to search for potential genomic targets of divergent selection underlying multiple fold differences in abdominal fatness, which differs by 3-fold between fat (FL) and lean (LL) lines or 12-fold between high growth (HG) and low growth (LG) lines. Briefly, eight DNA libraries were prepared for paired-end sequencing of pooled genomic DNA from multiple sires or multiple dams across the four genetic lines. The FL and LL (sire and dam) DNA libraries were sequenced in four separate lanes on an Illumina GA IIx, while the four libraries from the HG and LG were sequenced on a HiSeq 2000 sequencer. We achieved a mean coverage of ~10x and 20x per library, respectively, on the Illumina GA IIx and HiSeq 2000 sequencers. Sequence data from the sires and dams within each of the four genetic lines were pooled and mapped to the chicken reference genome (build 2.1) for SNP detection, using CLC bio Genomics Workbench software. The SNP allele frequency was estimated within each line using allele counts at detected SNP positions. To minimize the confounding effect of genetic drift, the data across 2 lines were combined and used for selection sweep analyses. Therefore, the SNP allele frequency was averaged across HG and FL (HGFL), since both lines are characterized by a high abdominal fat phenotype. Likewise, the SNP allele frequency was averaged across LG and LL (LGLL), since both lines are characterized by a low abdominal fat phenotype. We then estimated the FST parameter, as a measure of genetic differentiation between HGFL and LGLL in overlapping windows of 100 SNP along autosomes. We have identified several potential genomic targets of divergent selection and putative candidate genes subjected to divergent selection within these regions.