Cellular heterogeneity underlies the regulation of complex traits and diseases, making cell-resolved molecular characterization fundamental for advancing precision agriculture and human biomedicine. Rapid advances in single-cell omics technologies, together with the ever-evolving computational methods, have led to the establishment of large-scale single cell atlases across multiple species, providing unprecedented opportunities to systematically dissect cellular diversity and regulatory mechanisms underlying complex phenotypes. In this review, we summarize recent progresses in single-cell sequencing technologies, reference cell atlas projects, and the integration of single-cell omics with population/quantitative genetics, as well as discuss how artificial intelligence and predictive perturbation modelling are reshaping single-cell research. Finally, we highlight future directions toward population-scale single-cell omics, AI-driven virtual cells/tissues, and genotype–cell/tissue–trait frameworks, which are expected to transform biological discovery, enable better design of veterinary medicines and human therapeutics, as well as accelerate sustainable farmed animal improvement and precision medicine.
Primordial germ cells (PGCs) are highly valuable for chicken genetic engineering; however, their cultivation is frequently hindered by spontaneous cell aggregation and slow proliferation in certain breeds. We systematically evaluated the culture dynamics of PGCs derived from four distinct chicken breeds. To investigate the molecular mechanisms driving these phenotypes, we performed integrated transcriptomic and proteomic profiling, alongside targeted exogenous supplementation assays. We found that White Leghorn PGCs exhibited superior line establishment efficiency, whereas broiler breeds, particularly White Cornish, displayed severe cell aggregation and delayed growth. Transcriptomic and proteomic analyses revealed that WC PGCs exhibit an inherent deficiency in glutathione metabolism and an activation of the p53 signaling pathway. This finding was experimentally corroborated by the accumulation of reactive oxygen species (ROS) in these cells. Based on these mechanistic insights, we experimentally demonstrated that targeted exogenous supplementation with L-sulforaphane and ascorbic acid effectively mitigated excessive ROS and ameliorated the aggregation phenotype in White Cornish PGCs, with ascorbic acid notably accelerating cell proliferation. Subsequently, ascorbic acid also promoted proliferation in both Rhode Island Red and White Plymouth Rock lines, while alleviating aggregation in the White Plymouth Rock line. These findings demonstrate the cross-breed applicability of ascorbic acid, suggesting it could serve as a general supplement for PGC culture in other chicken breeds. Collectively, our study proposes a mechanistic model in which oxidative stress promotes PGC aggregation and offers practical antioxidant interventions to enhance the derivation efficiency of chicken PGCs, providing a valuable reference for the future optimization of in vitro culture media.
The rapid development of artificial intelligence (AI) is transforming modern society including animal genetics and breeding which is essential for meeting global nutritional needs in supplying milk, meat, and eggs among the most complete sources of protein and micronutrients. Although animal breeding has a long history of using data-driven prediction methods, the application of modern AI approaches remains at a relatively early stage and shows substantial potential for wide-ranging applications in animal genetics and breeding. This study aims to review the current use of AI tools in livestock species across three major applications of high interest: AI-assisted phenotyping, AI-enabled dissection of complex-trait biology (focusing on AI-assisted GWAS and AI-for-omics), and AI-assisted genomic prediction. In addition, we discussed emerging methods and tools developed in human genetics that are not yet widely adopted in farm animals but may offer translational potential for future livestock studies. Looking ahead, AI breakthroughs in livestock genetics and breeding are likely to be driven by increases in data scale and resolution, enhanced computational capacities, standardized data infrastructure and benchmarking, and systematic method development and validation. From a livestock genetic improvement perspective, the key value of AI will lie in integrating AI-generated inputs, including improved phenotyping and biological insights, into practical breeding decisions through statistical modeling.
Despite the astonishing growth of livestock single-cell transcriptome data, its full potential is severely limited by the lack of specialized functional databases and annotation tools. The ability to systematically integrate, standardize, and deeply interpret livestock single-cell data remains very limited. This study constructed SCAN, a high-quality resource platform integrating over 2.1 million single-cell datasets from six livestock species (Sus scrofa, Gallus gallus, Bos taurus, Bos grunniens, Ovis aries, and Oryctolagus cuniculus) and a representative model organism (Danio rerio). This dataset covers 33 tissue types and 155 cell types, providing differential gene expression profiles for each cell type. Based on this platform, we developed standardized analysis pipelines and visualization tools, and established a livestock marker gene library and a high-quality chicken single-cell annotation reference dataset (CARD). In addition, to demonstrate the utility of the platform, we used SCAN to perform a cross-species single-cell comparison of spermatogenesis between avian and mammalian species. The analysis showed that the early stages of spermatogenesis were relatively conserved across species, whereas differences increased at the later stages; chicken SPT and sperm were significantly enriched in catabolism-related pathways. These observations suggest that the late stages of chicken spermatogenesis may employ a metabolic strategy distinct from that of mammals, but the underlying mechanisms remain to be validated by functional experiments. This study provides a unified and standardized resource platform (SCAN) for livestock single-cell data, greatly facilitating data access, cell annotation, and downstream biological exploration, thereby enabling downstream biological discovery—as exemplified by the spermatogenesis study presented in this research. The SCAN database is freely and permanently available at http://www.scanimalsdb.bio-zz.cn .
Abstract Background The rumen microbiome is central to fiber degradation in dairy cows, yet the precise functional and metabolic networks driving variations in neutral detergent fiber (NDF) digestibility remain poorly understood. Elucidating these microbiome-host interactions is critical for improving forage utilization and advancing sustainable dairy production. Results Employing an integrative metagenomics-metabolomics approach, we characterized the rumen microecology of lactating dairy cows exhibiting high and low NDF digestibility phenotypes ( n = 10/group). Metagenomic profiling revealed a significant enrichment of fibrolytic taxa ( Prevotella sp., Lachnospiraceae bacterium) and carbohydrate-active enzymes in the high-NDF group (HI). In addition, species-level differential abundance testing showed that Methanobrevibacter spp. was more abundant in the low-NDF group (LO) than in the HI group, whereas no archaeal differences were detected at the phylum or genus levels. Functional and metabolomic analyses indicated broader enrichment of amino acid degradation functions and higher concentrations of major volatile fatty acids (VFAs) in HI cows, consistent with stronger overall ruminal fermentation activity. Furthermore, VFAs and microbial composition explained 54.3% and 43.0% of the variation in NDF digestibility, respectively. An exploratory machine learning model based on selected annotated metabolite features showed moderate discriminatory performance, with an out-of-bag accuracy of 77.19%. Conclusions Our findings suggest that coordinated shifts in rumen microbial composition, amino acid metabolism, and VFA production are associated with variation in NDF digestibility in dairy cattle. These results provide a basis for future studies aimed at improving fiber utilization through microbiome-informed nutritional or breeding strategies.
