Predicting dairy cattle PL via longitudinal rumen microbiome dynamics using machine learning approaches.

Area of Science:

• Microbiology
• Animal Science
• Genomics

Background:

• The gut microbiota plays a role in aging in mice and humans.
• The relationship between rumen microbiota and aging in dairy cows is not well understood.
• Dairy cow longevity is crucial for farm profitability, but average productive lifespan is low.

Purpose of the Study:

• To characterize rumen microbial differences across parities in dairy cows.
• To identify microbial markers associated with productive lifespan (PL) and farm profitability.
• To provide insights for improving PL and farm profitability through rumen microbiota modulation.

Main Methods:

• 16S rRNA amplicon sequencing of rumen microbiota from 341 dairy cows.
• Machine learning analysis (Support Vector Regression, Random Forest) to predict PL and profitability.
• SHapley Additive exPlanations (SHAP) analysis to identify key microbial markers.

Main Results:

• Rumen microbiota composition changes with increasing parity: decreased alpha diversity, weakened interactions, increased Proteobacteria, decreased Bacteroidetes.
• Machine learning models identified microbial markers predictive of PL (AUC=0.788) and farm profitability (AUC=0.763).
• The genera Eubacterium_hallii_group and Prevotella_7 were identified as key indicator strains for PL and profitability.

Conclusions:

• Alterations in the rumen microbiota are associated with aging in dairy cows.
• Microbial markers can predict productive lifespan and farm profitability.
• Modulating rumen microbiota presents opportunities to enhance dairy cow longevity and sustainable farming practices.