Supplementary materials

Heat stress poses a major threat to dairy cattle productivity, particularly in high-producing Holstein cows. To identify robust biomarkers of thermotolerance, we employed an integrative strategy combining physiological phenotyping, blood metabolite profiling, and transcriptomic analysis. A total of 120 lactating Holstein cows were evaluated under natural summer heat conditions using rectal temperature, respiratory rate, salivation index, serum HSP70, cortisol, potassium levels and milk production. These seven indicators were weighted via an entropy-based TOPSIS model to classify heat-resistant (HR) and heat-sensitive (HS) individuals. Subsequent transcriptomic analysis identified 330 differentially expressed genes (DEGs), with PDGFRA upregulated and TIMP3 and CCL5 downregulated in HR cows, suggesting reduced inflammation and extracellular matrix stress. Untargeted metabolomics revealed 220 differentially expressed metabolites, with HR cows exhibiting lower levels of anti-inflammatory compounds such as 6-Gingerol and PIP (18:1), and higher levels of inflammatory lipids. Two plasma metabolites, 3-Methoxytyramine (3-MT) and (3Z)-Phytochromobilin (PΦB), showed strong discriminative power for thermotolerance (AUC > 0.88). Multi-omics integration uncovered 411 significant gene–metabolite correlations enriched in heat-related pathways, including sphingolipid signaling and arachidonic acid metabolism. The identified biomarkers demonstrated their utility for rapid, noninvasive screening of heat-tolerant cows. These findings provide novel insights into the molecular mechanisms of heat resilience and offer a foundation for biomarker-assisted selection in climate-resilient dairy breeding.