Haplotype Purging After Relaxation of Selection in Lines of Chickens that Had Undergone Long-Term Selection for High and Low Body Weight

Abstract

Bi-directional selection for increased and decreased 56-day body weights (BW56) has been applied to two lines of White Plymouth Rock chickens-the Virginia high (HWS) and low (LWS) body weight lines. Correlated responses have been observed, including negative effects on traits related to fitness. Here, we use high and low body weight as proxies for fitness. On a genome-wide level, relaxed lines (HWR, LWR) bred from HWS and LWS purged some genetic variants in the selected lines. Whole-genome re-sequencing was here used to identify individual loci where alleles that accumulated during directional selection were purged when selection was relaxed. In total, 11 loci with significant purging signals were identified, five in the low (LW) and six in the high (HW) body weight lineages. Associations between purged haplotypes in these loci and BW56 were tested in an advanced intercross line (AIL). Two loci with purging signals and haplotype associations to BW56 are particularly interesting for further functional characterization, one locus on chromosome 6 in the LW covering the sour-taste receptor gene PKD2L1, a functional candidate gene for the decreased appetite observed in the LWS and a locus on chromosome 20 in the HW containing a skeletal muscle hypertrophy gene, DNTTIP1.

Keywords: Virginia chicken lines; advanced intercross line; body weight; directional selection; domestication; haplotype; purging; relaxed selection.

Figure 1

Genome-wide scans for regions with significant haplotype changes. (A) LW and (B) HW selected Virginia chicken lineages. Purging regions (see 3.2. below) are highlighted by black arrows.

Figure 2

Genome-wide and regional hapFLK trees identifying purging in the HW and LW lineages. Panels (A–C) show trees for the LW lineage, with (A) suggesting strong overall genome-wide purging and (B,C) illustrating the two strongest regional purging signals on chromosome 6. Panels (D,F) show the hapFLK trees for the HW lineage, with (D) illustrating that ongoing selection dominated on the genome-wide level and (E,F) illustrating the strongest purging signals on chromosomes 8 and 20. The trees for the remaining seven purging regions are provided in Figures S2–S9.

Figure 3

Purging regions in the LW lineage with associations to 56-day body weight. The figure presents results for the candidate purging region around 17Mb on chromosome 6. In (A), the large changes in frequencies consistent with purging were observed for haplotypes Hap1_{LC6_17} and Hap2_{LC6_17} across the four LW populations. In (B,C), box plots of 56-day body weights are shown for chickens with no, one, and two copies of these haplotypes (Purged Hap2_{LC6_17} in (B); Selected Hap1_{LC6_17} in (C). Overall p-values in (B)/(C) were from the one-way ANOVA analysis. Stars denote the significance of pair-wise t-tests between haplo-genotypes (***/*/ns represent p < 0.001/p < 0.05/non-significant).

Figure 4

Purging regions in the HW lineage with associations to 56-day body weight. The top panel presents results for the purging region on chromosome 20 (10.7–10.8 Mb). In (A), the large changes in haplotype frequencies for Hap1_{HC20_10} and Hap2_{HC20_10} across the four HW populations are consistent with purging. In (B,C), box plots of 56-day body weights are shown for chickens with no, one, and two copies of these haplotypes (Hap1_{HC20_10} in (B); Hap2_{HC20_10} in (C)). In bottom panel, (D) haplotype frequency changes consistent with purging are shown in the purging region on chromosome 8 (2.4–2.7 Mb). In (E,F), box plots illustrate 56-day body weight distributions for AIL individuals with no, one, and two copies of these haplotypes (Hap3_{HC8_2} in (B); Hap1_{HC8_2} in (C)). Overall p-values in B/C/E/F were from the one-way ANOVA analysis. Stars denote the significance of pair-wise tests between genotypes calculated by a t-test (*/**/ns represent p < 0.05/0.01/non-significant).