Functional gene expression is closely linked to an organism’s physiology and can be quantified using Real-Time Quantitative Reverse Transcription PCR (RT-qPCR). However, the stability of reference gene expression is not absolute, which may impact the accuracy of RT-qPCR results. In this study, we evaluated the suitability of nine genes including receptor for activated protein kinase c1 (rack1), ribosomal protein L6 (rpl6), ribosomal protein L9 (rpl9), ribosomal protein S2 (rps2), ribosomal protein S18 (rps18), ribosomal protein lateral stalk subunit P0 (rplp0), eukaryotic translation elongation factor 1β (eef1b), eukaryotic translation initiation factor 4a (eif4a), eukaryotic translation initiation factor 5a (eif5a) analyzed from RNA sequencing (RNA-Seq) data in addition to three genes including eukaryotic elongation factor 1α (eef1a), β-actin (actb), and glyceraldehyde 3-phosphate dehydrogenase (gapdh) selected from the literature to obtain the best internal controls in the RT-qPCR analysis of M. rosenbergii under overmating stress and natural aging. RefFinder was used to comprehensively evaluate the stability of the candidate reference genes. The initial results showed that three genes (eif5a, rps18, and rplp0) from the RNA-Seq data had relatively stable expression levels, which were more stable than those of the three commonly used reference genes. Eif5a and rps18 were the best combination for the RT-qPCR analysis of M. rosenbergii under overmating stress and aging. Further analysis indicated that eif5a might be the best reference gene for the study of M. rosenbergii.

Decapod Iridescent Virus 1 (DIV1) causes severe economic losses in Macrobrachium rosenbergii farming. Serum metabolic biomarkers are valuable tools for disease-resistant breeding, yet most existing studies employ mixed-sex samples, and the impact of sexual dimorphism on screening reliability remains poorly quantified. In this study, non-targeted serum metabolomics was performed on 28 healthy M. rosenbergii (14 females and 14 males), which were divided into marginally resistant and susceptible groups after DIV1 challenge. We compared results from mixed-sex and sex-specific screening strategies, and separately analyzed the overlap of upregulated and downregulated differential metabolites. Significant basal sexual dimorphism was observed in serum metabolism. Mixed-sex analysis identified 247 differential metabolites, while male-specific and female-specific analyses identified 121 and 161, respectively. Only 26 metabolites showed consistent directional changes across all three comparisons, accounting for 10.5% of total mixed-sex differential metabolites. Nearly 90% of mixed-sex results were either sex-specific signals or independent of sex. Analyses that do not account for sex introduce a large number of sex-unrelated differential markers and mask sex-specific metabolic features, and sex-specific screening is a necessary strategy to obtain reliable metabolic biomarkers for crustacean disease-resistant breeding.