Skip to main content

Genomic footprints of recovery in the European bison

Cite this dataset

Druet, Tom et al. (2020). Genomic footprints of recovery in the European bison [Dataset]. Dryad.


After extinction in the wild in the beginning of the twentieth century, the European bison has been successfully recovered in two distinct genetic lines from only twelve and seven captive founders. We here aimed at characterizing the levels of realized inbreeding in these two restored lines to provide empirical insights into the genomic footprints left by population recovery from a small number of founders. To that end, we genotyped 183 European bison born over the last 40 years with the Illumina BovineHD beadchip that contained 22,602 informative autosomal SNPs after data filtering. We then identified homozygous-by-descent (HBD) segments and classified them into different age-related classes relying on a model-based approach. As expected, we observed that the strong and recent founder effect experienced by the two lines resulted in very high levels of recent inbreeding and in the presence of long HBD tracks (up to 120 Mb). These long HBD tracks were associated with ancestors living approximately from 4 to 32 generations in the past suggesting that inbreeding accumulated over multiple generations after the bottleneck. The contribution to inbreeding of the most recent groups of ancestors were however found to be decreasing in both lines. In addition, comparison of Lowland individuals born at different time periods showed that the levels of inbreeding tended to stabilize, HBD segments being shorter in animals born more recently which indicates efficient control of inbreeding. Monitoring HBD segment lengths over generations may thus be viewed as a valuable genomic diagnostic tool for populations in conservation or recovery programs.

Usage notes

BBO3.ped: ped file with family, sample ID, father ID, mother ID, sex, phenotype, genotypes
For family, we indicated the genotyping batch:
WLl = samples genotyped in Poland; WLi = samples genotyped at Labogena (from INRA); WLC = sample genotyped at GIGA platform (University of Liege, Belgium)

All samples WLl and WLi are from the Lowland line except five sample with ID starting with "k" (lines 138 to 142), these are from the Lowland-Caucasian line.
All samples WLC are from the Lowland-Caucasian line.
Parents ID and sex are unknown and there is no phenotype. map file with chromosome, marker ID, genetic position (set to 0) and physical position (we used Bovine UMD3.1 assembly)

BBO3.yob: file with estimated birth year with ID, family ID (WLl, WLi, WLC), year of sampling, age at sampling (in months), estimated birth year.
The birth year was estimated by assuming sampling was one on July 1st (middle of the year)


University of Warmia and Mazury (Olsztyn, Poland), Award: n° 0105-804

F.R.S.-FNRS, Award: J.0134.16

F.R.S.-FNRS, Award: J.0154.18