Hematology RIs help clinicians and researchers determine whether a hematology parameter is abnormal, which plays an important role in animal health surveillance. China is one of the largest dairy producers in the world, with millions of Holstein cows. However, there has been no published data on hematology RIs for dairy cows in China yet. Therefore, the aim of this study is to establish updated and accurate RIs for Holstein cows in southern China. To increase the accuracy of the RIs, we collected blood samples from 786 Holstein cows and analyzed 25 hematologic variables. The RIs for Holstein cows were established using the 95% percentile RIs according to the American Society of Veterinary Clinical Pathology guidelines. The effects of different ages, parities and lactation stages were also checked in this study. The data of 21, 22 and 19 out of 25 hematology parameters were significantly different between different ages, parities and lactation stages, respectively. Furthermore, the hematology RIs of separate subclasses according to different ages, parities and lactation stages were generated. Hematology RIs according to ages and lactation stages, as well as parities and lactation stages, were also assessed. Together, our results confirm that hematology RIs for cows vary by ages, parities and lactation stages. The present study helps to fill the gap in hematology RIs for Holstein cows in southern China, and our data may serve as a very useful tool for monitoring the health and welfare of dairy cattle in China.

Mastitis is a common disease for dairy cows that exerts tremendously detrimental impacts on the productivity of cows and economic viability of pasture. Dihydromyricetin (DMY) is a flavonoid monomeric compound that possesses anti-inflammatory and antioxidant activity. This study aimed at dissecting the effects of DMY on the lactation performance, blood parameters, gut microbiota, and metabolite profiles of dairy cows with subclinical mastitis (SM). The results showed that dietary supplementation with DMY resulted in a reduction in milk somatic cell count, an increase in serum T-AOC and CAT activity, as well as a decrease in serum MDA content. DMY significantly enhanced the prevalence of Coprococcus and Roseburia and reduced the proportion of Cyanobacteria, Proteobacteria, and Dehalobacterium. The amino acid degradation, antibiotic resistance, and O-antigen building blocks biosynthesis (E. coli) capacity of gut microbes were notably diminished by DMY supplementation in cows with SM. Moreover, fecal and plasma metabolomic analysis revealed that DMY intervention reduced the abundance of pro-inflammatory metabolites including arachidonic acid analogues, ω-6 PUFA, and structural components of bacteria. Nevertheless, the levels of anti-inflammatory and antioxidant metabolites involving secondary bile acids, antioxidant vitamins, specific amino acid analogues, etc. were elevated by DMY administration. Overall, DMY might ameliorate SM via enhancing antioxidant capacity and improving the structure of the hindgut microbial community and metabolite profiles in dairy cows. These findings underscore the potential of DMY as a valuable dietary supplement for the improvement of mammary inflammatory diseases in dairy cows.

High-yielding dairy cows suffer from a high metabolic load and oxidative stress, which lead to systemic inflammation and metabolic disorders, increasing the susceptibility of these cows to various production diseases. Dihydromyricetin (DMY) has demonstrated potent antioxidant and anti-inflammatory physiological functions; however, research into its application in ruminants remains limited. This study investigated whether DMY supplementation is associated with the maintenance of metabolic homeostasis through the regulation of gut microbiota and metabolite profiles. A total of 14 mid-lactation Holstein dairy cows were randomly divided into two groups (n = 7 per group) and supplemented with DMY at 0 or 0.05% in their basal diet for 60 consecutive days. The effects of DMY on the blood biochemical indicators and the antioxidant capacity of the dairy cows were then determined. Alterations to the gut microbiome and the fecal and plasma metabolome were analyzed through 16S rDNA sequencing and untargeted metabolomics. The results showed that DMY significantly improved the activity of serum glutathione peroxidase (GSH-Px) and presented a trend of increasing the total antioxidant capacity (T-AOC). The abundance of multiple fiber-degrading and beneficial commensal bacteria in the gut, including Fibrobacter_succinogenes, Ruminococcus_albus, and Turicibacter, was significantly elevated by the DMY intervention, accompanied by the upregulation of 8,11,14-eicosatrienoic acid, myricetin, dihydro-3-coumaric acid, PGE1, L-leucine, nicotinuric acid, pantothenic acid, and pyruvate in the feces and plasma. Moreover, DMY supplementation notably reduced the abundance of potential pathogenic microbes, such as Chloroflexi, Deltaproteobacteria, RFP12, and Succinivibrio, and downregulated the levels of 12-hydroxydodecanoic acid, 12,13-DHOME (12,13-dihydroxy-9Z-octadecenoic acid), 16-hydroxyhexadecanoic acid, niacin, and glycerol 3-phosphate. These differential metabolites were principally enriched in the mTOR signaling pathway; pantothenate, nicotinate, and thiamine metabolism; glutathione metabolism; and glycolysis/gluconeogenesis. In summary, dietary supplementation with DMY increased the abundance of intestinal fiber-degrading bacteria and multiple metabolites with known anti-inflammatory and antioxidant properties in the feces and plasma, and was associated with alterations in metabolic pathways involving B-vitamins, amino acids, and glutathione. This suggests a potential role for DMY in supporting metabolic homeostasis in dairy cows.

Genome-wide association studies (GWAS) have been a successful tool for identifying quantitative trait loci (QTL) for economically important traits in dairy cows. However, the availability of QTLs linked to phenotypic traits is limited in the literature. In this study, we used GWAS, haplotype association, and fine-mapping analyses to identify candidate variants associated with milk production, health, and reproductive traits in 380 Chinese Holstein cattle from Southern China using whole-genome sequence data. GWAS identified 91 genome-wide significant signals that were annotated to 63 genes associated with milk production, health, and reproductive traits in dairy cattle. Haplotype association analysis further revealed that eight GWAS signals within three QTLs were associated with milk production and health traits of cows. Fine-mapping analysis revealed that 3 GWAS signals (6_92530313_G_A, 10_17185230_G_A, and 10_17209112_T_G) were the potential causal variants. Several candidate genes, including ANKS1B, IL17RD, CNOT6L, AOC1, and TLE3, have been confirmed to be associated with milk production, health, and reproductive traits in dairy cows. These findings significantly contribute to unraveling the genetic basis of economically important traits in Holstein cattle.

Neural additive models have demonstrated strong flexibility and interpretability for nonlinear data analysis in various scenarios, such as medical data analysis. However, most existing methods implicitly assume Gaussian stochastic perturbations, an assumption that is too strict in real-world time-series applications where heavy-tailed noise, abrupt outliers, and asymmetric error structures are common. These limitations can substantially degrade predictive performance and undermine feature-level interpretability. To overcome these limitations, we introduce a S parse N eural H uber A dditive M odel (SpNHAM) tailored for robust forecasting under non-Gaussian noise. SpNHAM combines Huber-based adaptive loss reweighting with a structured sparsity penalty to simultaneously improve robustness, promote component-wise interpretability, and enable automatic feature selection in multivariate temporal settings. The model is efficiently optimized using SGD on the differentiable approximation obtained via Nesterov’s smoothing, making it scalable to high-dimensional time series. Extensive experiments on synthetic datasets and real-world benchmarks demonstrate that SpNHAM consistently outperforms existing neural additive and robust forecasting baselines, particularly in the presence of contaminated noise, while maintaining promising interpretability.