Targeted Intervention Strategies for Maternal–Offspring Transmission of Christensenellaceae in Pigs via a Deep Learning Model

Abstract

Understanding the mechanisms of maternal microbial transmission is crucial for early gut microbiota development and long-term health outcomes in offspring. However, early maternal microbial interventions remain a challenge due to the complexity of accurately identifying transmitted taxa. Here, the maternal–offspring microbial transmission model (MOMTM), a deep learning framework specifically designed to map maternal microbiota transmission dynamics across pig breeds and developmental stages, is introduced. Using MOMTM, key transmitted taxa, such as the Christensenellaceae R-7 are successfully predicted, which show high transmission centrality during early development periods. Additionally, it is demonstrated that galacto-oligosaccharide intervention in sows promotes a Christensenellaceae R-7-dominated enterotype and improves fiber digestibility in offspring. Further analysis reveals that Christensenellaceae, particularly Christensenella minuta, have enhanced adhesion and mucin utilization capabilities, facilitating its gut colonization. These findings highlight MOMTM's potential as a novel approach for microbiota-targeted health interventions in early life, offering insights into strategies that promote gut health and development from birth.