On the path to recovery : characterizing inbreeding in the European Bison with a model-based approach

When two alleles from the same individual descend from a single allele in an ancestor, they form an homozygous-by-descent (HBD) or autozygous segment. Increased autozygosity is frequently associated with negative effects on fitness. Population bottlenecks or founder effects can dramatically increase autozygosity levels. We recently developped a model-based approach to identify HBD segments that is particularly valuable when marker information is sparser (e.g., low-fold sequencing experiments, low marker density, marker recruitment bias). This hidden Markov model is also useful when the data is not uniform (variable genotyping errors, marker spacing or recombinations rates). Overall the approach is beneficial in situations encountered in wild organisms. Here, we use our model to characterize individual autozygosity in the restored European Bison. After complete extinction in the wild, restoration programs started as soon as 1929 using only twelve founders kept in European zoos and private breeding centers and leading to the sucessfull restoration of two genetic lines, the lowland and lowland-Caucasian lines. To characterize inbreeding we used a bovine genotyping array encompassing more than 700,000 SNPs. Only 22,602 SNPs were conserved after data filtering. Despite the sparse marker information, we estimated high inbreeding levels (0.30 to 0.40) consistent with pedigree estimates. Many long HBD segments (up to 120 Mb) associated with recent common ancestors, approximately 4 to 32 generations in the past, were identified. Their distribution is compatible with a recent bottleneck. The results obtained with this sparse genotyping data were validated with whole genome sequence data from two individuals. Overall, the presented approach is still efficient when the number of informative markers is reduced compared to traditional high-density genotyping arrays and can also be used with genotyping-by-sequencing data. The same strategy can be applied to wild populations for which the history is not well documented