Abstract
Inbreeding can have negative effects, such as increasing the expression of deleterious alleles or reducing fitness. A method based on Hidden Markov Models (HMM) was developed to determine the probability of an individual genome in a homozygous-by-descent state (HBD). As a result of an experiment of divergent selection for birth weight environmental variability two lines were created: high variability line (H-Line) and low variability line (L-Line). The L-Line demonstrated a better performance in traits related with robustness than the H-Line. From a selection period of 20 generations, a total of 655 individuals from the H-Line and 675 individuals from the L-Line were genotyped with a high-density SNP array. We used a predefined multiclass HMM with a total of 9 age related HBD classes and 1 non HBD class. The sum of the probabilities of each HBD class was defined as the total HBD inbreeding (F(HBD)). In addition, F(HBD) was divided into age related groups as recent and ancient. Moreover, recent pedigree inbreeding (F(PEDR)) was defined using different generation thresholds (4 to 14). The evolution of F(HBD) across generations was similar in both selected lines. However, the distribution in each age-related class was different between lines in more recent generations. The H-Line presented twice as much F(HBD) by ancestors from 8 generations ago than the L-Line. Moreover, the correlations between recent F(HBD) and F(PEDR) obtained with different generation thresholds were greater in the H-Line when very recent F(HBD) was calculated from classes related with ancestors from 1 to 8 generations ago. However, in the L-Line, considering more than 4 generations ago to define very recent inbreeding did not affect the correlations with F(PEDR). The HBD was the first methodology that could detect differences in the inbreeding pattern between the selected lines that could be related with the divergent selection, despite being under the identical mating policy and similar intensity of selection.