Transcriptomic Analysis and Machine Learning Identify Cross-Pathogen Biomarkers for Bacterial and Parasitic Infections in Silver Pomfret (Pampus argenteus)

Silver Pomfret is increasingly threatened by many diseases under intensive artificial culturing conditions, yet conserved host biomarkers across different infections remain poorly defined. In this study, we integrated transcriptomic datasets from independent infections with Cryptocaryon irritans, Nocardia seriolae, and Photobacterium damselae subsp. damselae to identify shared host-response genes. By combining differential expression analysis with weighted gene co-expression network analysis, we prioritized six candidate genes associated with cross-pathogen infection responses. Random Forest and support vector machine analysis further supported their classification potential across the three infection models. Phylogenetic and structural analyses provided additional evidence for the conserved annotation of these proteins. GSVA-based signature analysis supported the cross-pathogen discriminatory capacity of the six-gene panel and suggested context-dependent contributions of individual genes across infection models. Immune signature analysis indicated distinct host immune response patterns under different pathogenic challenges, and candidate genes showed positive associations with inferred T cell-related signatures. Upstream regulatory prediction identified CTCF and the miR-17/20/93 family as potential regulators of these genes. Quantitative real-time PCR of the kidney further highlighted canx, rnd3, and angptl4 as the most robust infection-responsive candidates, with consistent temporal expression patterns observed from 0 to 24 h post-infection. These findings suggest a potential cross-pathogen host-response pattern in Silver Pomfret and provide preliminary support for future exploration of molecular markers for disease monitoring in aquaculture.