Animal Science Department


Date of this Version



2023 Poultry Science.


Open access.


The population dynamics of 2 lines of chickens from a long-term (59 generations) selection experiment were assessed based on pedigree data. These lines were propagated from phenotypic selection for low and high 8-wk BW in White Plymouth Rock chickens. Our objective was to determine whether the 2 lines maintained similar population structures over the selection horizon to allow meaningful comparisons of their performance data. A complete pedigree of 31,909 individuals, consisting of 102 founders, 1,064 from the parental generation, and 16,245 low weight (LWS) and 14,498 high weight (HWS) select chickens, was available. Inbreeding (F) and average relatedness (AR) coefficients were computed. Average F per generation and AR coefficients were 1.3 (SD 0.8) % and 0.53 (SD 0.001) for LWS, and 1.5 (SD 1.1) % and 0.66 (SD 0.001) for HWS. Mean F for the entire pedigree was 0.26 (0.16) and 0.33 (0.19), and maximum F was 0.64 and 0.63, in LWS and HWS, respectively. Based on Wright’s fixation index, at generation 59, substantial genetic differences were established between lines. The effective population size was 39 in LWS and 33 in HWS. The effective number of founders was 17 and 15, effective number of ancestors were 12 and 8, and genome equivalents were 2.5 and 1.9 in LWS and HWS, respectively. About 30 founders explained the marginal contribution to both lines. By generation 59, only 7 male and 6 female founders contributed to both lines. Moderately high levels of inbreeding and low effective population sizes were inevitable, as this was a closed population. However, effects on the fitness of the population were expected to be less substantial because founders were a combination of 7 lines. The effective numbers of founders and ancestors were relatively low compared to the actual number of founders, as few ancestors contributed to descendants. Based on these evaluations, it can be inferred that LWS and HWS had similar population structures. Comparisons of selection responses in the 2 lines therefore should be reliable.