Title: Multi-trait breeding value-based selection and prediction of genetic gain under alternative breeding scenarios in common carp (Cyprinus carpio).

Cyprinus carpio, popularly known as common carp is the fourth most farmed fish in the world. This fish is an exotic introduction in India with limited founding stock but is now very common in water bodies. Long history of inbreeding and negative selection has negatively impacted its aquaculture production. To revive its population and improve productivity, ICAR-CIFE, Mumbai, India, has initiated a selective breeding program targeted at an inland saline production environment. Barren lands challenged with soil salinity and underground saline water are a potential resource for common carp aquaculture owing to its ability to tolerate low saline conditions. The breeding goal is to develop a fast-growing and resilient strain of common carp suitable for inland saline aquaculture systems. In the present study, the utilised growth data comprised a multi-trait, multi-environment dataset. A multi-trait Best Linear Unbiased Prediction (BLUP) approach was employed. Fish reared under low salinity (LS; 2–4 PSU) exhibited superior growth performance, with mean body weight (BW), standard length (SL), and body depth (BH) being higher by 156 g, 1.3 cm, and 0.6 cm, respectively, than those cultured under high salinity (HS; 6–8 PSU) conditions. The growth traits exhibited high heritability (0.36–0.57) and displayed positive between-salinity genetic correlations between 0.76 and 0.97, indicating a low genotype-by-environment interaction. The full-sib effect correlation was moderate to high across traits, indicating a low full-sib by salinity interaction. The between-trait pond effect correlation within each salinity was found to be high (0.80-1.00). A relatively high accuracy (0.72) of multi-trait BLUP (aggregate genotype) was estimated, indicating that a reliable multi-trait selection can be achieved. The theoretical response to aggregate genotype selection was +4.25 index units. Further, a stochastic simulation was performed to delineate the most effective breeding. Specifically, four alternative breeding scenarios, varying in the number of families produced and selected in each generation, were simulated for 10 generations with 10 replicates each. The true breeding values (TBV), BLUP, accuracy, and genetic gain for multi-traits were predicted. Breeding scenario 2 was selected and included 125 families produced by nested mating and 20 fish evaluated per family (5000 total) per generation, followed by selection (maximum four per family) with 200 top males and 250 females selected and mated to produce the next generation. This scenario yielded a higher trend for TBV, accuracy, genetic gain, and a lower rate of inbreeding (0.5 %) than the alternative scenarios. The results of the present study will aid in multi-trait selection and the adoption of an effective breeding strategy for the breeding of common carp in India.