Abstract
IN BRIEF: Vaginal microbiota composition influences female fertility, however it has not been studied for measuring fertility level in female pigs. This study reveals significant vaginal microbiota composition differences between high reproductive performance and infertile sows, and demonstrates that the vaginal microbiota has promise for improving female pig selection using machine learning modeling. ABSTRACT: There is a need for reliable and effective biomarkers of female fertility and reproductive potential in the pork industry, as current selection protocols are not keeping up with the rate of improvement for other production-related traits. This study aimed to investigate the vaginal microbiota composition between sows of differing fertility status and identify candidate vaginal microbiota biomarkers of sow fertility. The vaginal microbiota of high reproductive performance sows (HRP, n = 52) with number of piglets born alive ≥13 and infertile sows (INF, n = 23), that remained nonpregnant after two consecutive rounds of artificial insemination, were investigated. Sequencing results revealed significantly different (P < 0.05) beta diversity at the genus level between HRP and INF vaginal microbiota communities. Accordingly, the composition of the vaginal microbiota diverged between HRP and INF sows, with INF sows having increased (P < 0.05) relative abundance of Lachnospiraceae XPB1014 group and HRP sows having increased (P < 0.05) relative abundance of Aerococcus and Staphylococcus at the genus level. Forty-two genera were selected as candidate biomarkers of sow fertility via partial least squares discriminant analysis (PLS-DA) and recursive feature elimination. The support-vector machine model classified sow fertility with 93.3% accuracy, supporting potential industry application to improve upon current methods for selection and recruitment in the breeding herd. Future investigations should validate the candidate vaginal microbiota biomarkers in a large, independent population of sows and gilts to evaluate their application for predicting future reproductive performance and assess their true industry applicability.