A) Principal component analysis with Beefalo (blue), bison-cattle hybrids (orange), and a bison-cattle F1 hybrid (black star) projected onto axes computed from bison (red) and taurine (yellow) and indicine (gray) cattle. PC1 separates bison and cattle while PC2 separates the two cattle groups. B) Comparison of Beefalo reported bison ancestry and position on PC1. Placement of bison hybrids on PC1 is correlated with reported bison ancestry, though Beefalo do not follow this trend. C) depicts the same as in B) though compares inferred bison proportions from ADMIXTURE to reported bison ancestry. Reported and inferred bison proportions match well for bison hybrids, but all Beefalo were inferred to have less bison ancestry than reported. D) Autosomal ADMIXTURE analysis of bison, taurine and indicine cattle, bison hybrids, and Beefalo. The position of Joe’s Pride (JP) is depicted with an arrow in A-D.

Allele sharing statistics using individual bison hybrids and Beefalo. Individuals are shown on the Y axes. Beefalo are shown in blue and bison-cattle hybrids in orange, with a bison-cattle F1 shown in black. Error bars depict 3 standard errors. JP = Joe’s Pride. A) Testing for the presence of bison ancestry using statistics of the form D(Beefalo, taurine cattle; bison, water buffalo). B) Quantification of the proportion of bison ancestry using f4-ratios. All bison hybrids have at least 50% bison ancestry, while only 8 Beefalo are inferred to have bison ancestry, ranging from 2-18%. Reported bison ancestry is shown in red. C) Testing for allele sharing between Beefalo and indicine cattle, relative to other taurine cattle. Many individual Beefalo are inferred to have significant allele sharing with indicine cattle. All Beefalo which were inferred to have bison ancestry in panels A) and B) display significantly negative values (due to the presence of bison ancestry, which is an outgroup to both cattle breeds), showing that any inferred indicine ancestry is not due to the presence of bison ancestry. For all panels, error bars depict 3 standard deviations.

Local ancestry inference of Beefalo and bison hybrid genomes, using bison and taurine and indicine cattle as potential sources. Inferred ancestry across the all autosomes are shown for A) a bison hybrid with reported 75% bison ancestry, B) a Beefalo with 37% reported, and 12% detected, bison ancestry, and C) Joe’s Pride, a foundational Beefalo with 37.5% reported, and no detected, bison ancestry. Ancestry along the genome is colored by inferred source: homozygous bison, taurine, and zebu ancestry are shown in gray, green, and blue, respectively, with heterozygous bison-taurine, bison-zebu, and taurine-zebu ancestry shown in yellow, red, and light blue. Panel D) shows the inferred global bison and indicine ancestry proportions for each bison hybrid and Beefalo individual, with bison ancestry proportion shown in gray and indicine proportion shown in blue. Three bison hybrids are inferred to have ∼75% bison ancestry, while eight Beefalo have detectable bison ancestry, ranging from 2-18%. Indicine ancestry is detected in most Beefalo at variable levels, ranging from 2-38%, with most Beefalo having between 2-18%.

A) ADMIXTURE analysis of bison, taurine and indicine cattle, bison hybrid, and Beefalo X chromosomes. Beefalo are largely inferred to have X chromosomes derived entirely from taurine cattle, though variable amounts of indicine and bison ancestry are also present in some Beefalo. The position of Joe’s Pride (JP) is indicated with an arrow. Bison hybrids have majority bison ancestry. B) Bovid Y chromosomal phylogeny. The indicine (zebu) clade is collapsed, as are those containing yak and wisent. Water buffalo (not shown) was used as an outgroup. Beefalo all have taurine Y chromosomes, while bison hybrids have bison Y chromosomes. C) Comparison of the reported (red) and ADMIXTURE inferred bison ancestry proportion on the autosomes (blue) and X chromosome (green) for the three bison hybrids and eight bison which had detectable autosomal bison ancestry.

Supervised ADMIXTURE modeling of Beefalo and bison hybrid ancestry, using panels of bison and taurine and zebu cattle.

Local ancestry inference results from a F1 bison-taurine cattle hybrid. Ancestry was correctly inferred to be almost completely heterozygous for bison and taurine ancestry across the genome, though with some small segments incorrectly assigned to be heterozygous for bison-zebu bison, likely because of the low divergence between cattle subspecies. These segments are all small and are increasingly removed with more stringent filtering (for window size and minimum SNPs to call a window, shown in panel titles).

Local ancestry inference results from a Brangus individual, showing the presence of both taurine and zebu (indicine) ancestry.

Comparison of local ancestry inference using either downsampled (∼2x) read data (A, D, J, M, P, S), high coverage (30-42x) read data (B, E, H, K, N, Q, T), or called genotypes (C, F, L, O, R, U) on seven individual Beefalo and bison hybrids which were sequenced to high coverage: 5851 (A-C), 5854 (D-F), 5861 (G-I), 5887 (J-L), 5895 (M-O), 5872 (P-R), 5874 (S-U).