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

BMC Bioinformatics

Table 6 Percent polar bear for each brown bear in this study, calculated via \( \widehat{f} \), AD-LIBS, and HAPMIX, if available

Bear	Origin	\( \widehat{f} \)	AD-LIBS	AD-LIBS conservative	HAPMIX
ABC01	Baranof Island, AK	8.63%	18.6%	15.5%	14.4%
ABC02	Baranof Island, AK	8.87%	18.8%	15.7%	14.8%
ABC03	Chichagof Island, AK	9.63%	19.4%	16.2%	N/A
ABC04	Chichagof Island, AK	9.03%	19.0%	15.8%	N/A
ABC05	Chichagof Island, AK	8.93%	19.1%	15.9%	N/A
ABC06	Admiralty Island, AK	6.56%	17.1%	14.2%	N/A
Adm1	Admiralty Island, AK	6.12%	16.6%	13.8%	N/A
Adm2	Admirality Island, AK	6.05%	17.0%	14.2%	12.6%
Bar	Baranof Island, AK	8.14%	18.5%	15.4%	14.3%
Chi1	Chichagof Island, AK	8.57%	18.6%	15.5%	N/A
Chi2	Chichagof Island, AK	8.69%	18.7%	15.6%	N/A
Den	Denali Natl. Park, AK	7.02%	14.5%	11.9%	N/A
GP01	Glacier Park, Montana	4.37%	17.2%	14.3%	N/A
GRZ	Kenai Peninsula, AK	3.30%	13.0%	10.7%	N/A
OFS01	Östanvik, Sweden	0.464%	5.35%	4.41%	N/A
RF01	Ruokolahti, Finland	0.319%	6.90%	5.67%	N/A
SJS01	Slakka, Sweden	0.211%	5.27%	4.33%	N/A
Swe	Dalarna, Sweden	0%*	4.89%	4.02%	N/A

The asterisk indicates that Swe was used as a model unadmixed brown bear in \( \widehat{f} \) calculations, making polar bear ancestry undetectable. HAPMIX was only run on the four ABC Islands brown bears with minimum 20X coverage, to ensure that heterozygous variant calls were reliable. The “AD-LIBS conservative” column shows AD-LIBS estimates corrected according to the percent of homozygous and heterozygous polar bear ancestry calls that were incorrect in simulations under the single-pulse model (Table 1)

ISSN: 1471-2105