Identification of transcriptome markers in blue catfish cryopreserved sperm to predict male reproductive performance for catfish aquaculture

Catfish is the most important species in US aquaculture, accounting for 70% of freshwater production. The production of hybrids of female channel catfish (Ictalurus punctatus) and male blue catfish (I. furcatus) constitutes over 50% of the total harvest due to their superior production traits and enhanced disease resistance. However, hybrids cannot be produced naturally, and male must be euthanized for sperm collection. An appropriate storage condition is essential to preserve the sperm's ability to fertilize eggs during the female spawning season. Cryopreservation is a widely used method for sperm storage. However, it has been shown to affect sperm gene expression in many vertebrate species. A high degree of individual variability among cryopreserved sperm was reported, resulting in huge variations in hatch rate. Since high-quality gametes are the prerequisites for hybrid catfish reproduction, in this study, we investigate the molecular mechanisms underlying male reproductive performance, which is critical for the accurate prediction of hatching rate and offspring performance in hatchery environments. Overall design: The sperm samples from 30 male blue catfish have been collected and divided into tow groups, the fresh one, which has been directly stored in -80? freezer, and the cryopreserved one, which exprerienced the cryopreserved procedures. And then the cryopreserved sperm samples have been thawed for fertilization. The embryo harch success rate and viability rate (40 days post-hatch) were documented. RNA was extracted from each sample and libraries were constructed, followed by the RNA sequencing. Differentially expressed genes (DEGs) were identified between the cryopreserved samples with high hatch rate and low hatch rate. The correlation between the expression level of DEGs and the hatch success rates were assessed to determine the potential gene markers which can serve as the predictors for hatch performance. Futhermore, a machine learning model was developed integrating the gene markers and physiological parameters to enable accurate prediction of hatch success rate.