Data-driven enterosignatures link gut microbiome reorganization to heat stress responses in lactating sows

Background: Heat stress (HS) can disrupt the gut microbiome, yet most livestock studies rely on taxonomic summaries that overlook the ecological structure of microbial communities. Enterosignatures (ES) as latent, co-occurring microbial assemblages learned from metagenomic data, offer a framework to capture these dynamics but have scarcely been applied in livestock HS research. Methods: = 12), which were divergently selected based on genomic breeding values (GEBVs) for heat tolerance, and exposed to HS conditions. Results were decomposed using non-negative matrix factorization (NMF), yielding 8 taxonomic (T-ES) and 5 functional (F-ES) subcommunities. Functional profiles (based on KEGG Orthology, KOs) were mapped to metagenome-assembled genomes (MAGs) to integrate metabolic attributes within each ES. Results: = 0.638) suggest possible functional restructuring warranting investigation in larger cohorts. Conclusion: This work presents the first use of ES to track microbiome responses to HS in lactating sows. ES revealed latent taxonomic and functional subcommunities with clear temporal reorganization, offering insights not detectable with standard clustering or diversity metrics. Although genetic-line effects were modest, several ES showed biologically relevant shifts, supporting ES as a hypothesis-generating exploratory framework for linking microbial ecology to physiological adaptation under HS conditions, while warranting validation in larger, controlled trials.