Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: AD-LIBS: inferring ancestry across hybrid genomes using low-coverage sequence data

Fig. 2

Accuracy of AD-LIBS estimates of the overall extent of polar bear ancestry, using simulated data. a and c refer to simulations with a single polar-brown bear admixture event 12,000 years ago, followed by isolation (single-pulse model); b and d refer to simulations in which a brown-polar bear admixture event 12,000 years ago is followed by continual breeding with unadmixed brown bears (migration model). All AD-LIBS runs in this figure used 5 reference individuals per ancestral population and 10kb windows. a and b: Inferred percent polar bear ancestry using AD-LIBS and \( \widehat{f} \) versus true percent polar bear ancestry. c and d: inferred percent polar bear ancestry of each type, according to AD-LIBS, versus true percent polar bear ancestry of each type. Each point represents the percent of a single simulated hybrid bear genome with a specific type of ancestry. e: overall accuracy of AD-LIBS inferences versus true percent polar bear ancestry, including both types of simulations. The line of best fit by least-squares regression is also shown. Accuracy decreases slightly as polar bear ancestry increases, probably due to the tendency of AD-LIBS to overestimate the extent of heterozygous ancestry (c and d)

Back to article page