Milk production efficiency: candidate genes, digestibility and methane emission for late lactation in a sample of Italian Holstein cows

The aim of this study was to investigate physiological and genomic differences between high- (HEFF) and low-efficiency (LEFF) Holstein cows, using a residual-based definition of productive efficiency derived from automatic milking systems (AMS) data. Individual efficiency was quantified as the deviation from expected daily milk yield, estimated using a linear mixed-effects model applied to AMS records. This approach enabled the identification of 68 HEFF and 68 LEFF cows within a single herd managed under uniform feeding conditions. A genome-wide association study (GWAS) based on selective genotyping using the GGP Bovine 100k SNP array identified 28 significant SNPs across 13 autosomes, mapping to 51 candidate genes. These genes included loci associated with lipid metabolism and energy utilisation (CD36, LPGAT1), methane-related efficiency (VPS13B), and immune regulation (TLR9), suggesting a complex genetic architecture underlying productive efficiency. To support the biological relevance of the residual-based classification, zootechnical parameters and apparent total tract digestibility (aTTD) were evaluated over three consecutive days in a subset of 15 HEFF and 15 LEFF cows during late lactation. HEFF cows produced more milk than LEFF (33.76 vs. 28.19 kg/d; p < 0.05) and emitted significantly less methane per litre of milk (14.40 vs. 17.20 g/L; p = 0.02), despite similar total methane output. Furthermore, HEFF cows exhibited higher fibre digestibility, particularly for neutral detergent fibre (aTTD NDF: 57.62 vs. 48.09%; p = 0.003). Overall, efficiency defined by model residuals corresponds to measurable biological differences in digestibility and methane emission intensity, supporting the use of residual-based phenotypes as practical indicators of productive efficiency in dairy cattle.