Sperm methylome profiling revealed DNA methylation variations in Piedmontese bulls with different reproductive performances

Artificial insemination (AI) has revolutionized the breeding of livestock species, becoming the most used method in bovine reproduction. Nowadays, early prediction of bull fertility is one of the main challenges for AI-centers as the entire process of bull breeder selection is considered time and money-consuming. It is based on animal selection (genotypic and phenotypic traits), field trial inseminations, and conventional laboratory semen analysis. However, notwithstanding the advance of genomic selection allowed reliable identification of genetically elite bulls, selected bulls with apparently normal semen quality can vary significantly in their field fertility. This suggests that conventional semen evaluation methods are not fully foolproof and other sperm characteristics need to be investigated to explain the remaining variation. Emerging data about abnormal sperm DNA methylation and infertility suggest that epigenetic traits are promising candidates for this purpose [1,2]. Therefore, this study tested the hypothesis that sperm-specific epigenetic characteristics, such as DNA methylation, could discriminate Piedmontese bulls with different reproductive performances. Using a genome-wide approach we have characterized sperm methylome from normal (NF) and high fertility (HF) Piedmontese bulls, by reduced representation bisulfite sequencing (RRBS). Bulls (20) were assigned to either the NF or HF group (10/group) based on an adjusted fertility index provided by the Piedmontese AI center. Reproductive efficiency was predicted by a direct measure of field fertility, calculated as the ability of bulls to make cows pregnant through AI corrected for a wide range of factors (year, season, female genetic, farming). NF fertility index ranged from 0.531 to 0.631 while HF bulls showed values between 0.973 and 0.706. As expected, Computer Assisted Sperm Analysis confirmed that motility parameters were comparable within groups (63,7% NF vs 67,4% HF). RRBS identified a total of 348889 cytosines (10X coverage in all samples) able in part to separate NF or HF animals. We identified 968 differentially methylated cytosine DMCs (FDR