Genomic prediction of survival against Vibrio harveyi in leopard coral grouper (Plectropomus leopardus) using GBLUP, weighted GBLUP, and BayesCπ

Abstract

Leopard coral grouper (Plectropomus leopardus) is a marine economic fish that widely inhabits the western Pacific. In recent years, the quantity of wild-captured leopard coral grouper has had difficulty meeting commercial and recreational needs. Therefore, it is necessary to develop the artificial culture of leopard coral grouper. However, disease outbreaks caused by Vibrio harveyi have hindered the sustainable development of the leopard coral grouper farming industry. The primary purposes of this study are to 1) estimate the genomic heritability of survival against V. harveyi in P. leopardus; 2) perform a genome-wide association study (GWAS) to determine the genetic architecture; and 3) compare the predictive accuracy variation of genomic methods using different density single nucleotide polymorphism (SNP) panels and cross-validation grouping strategies. We collected 800 leopard coral groupers from two commercial aquatic companies and challenged them using V. harveyi by abdominal injection to obtain the survival phenotypes. After resequencing and quality control, a total of 798 fish and 1,211,259 high-quality SNPs were retained for downstream analysis. The genomic heritability (underlying scale) of survival against V. harveyi in leopard coral grouper ranged from 0.16 to 0.24. Eight trait-related regions with genetic variance explained >1% were identified from the GWAS results, and thirty genes were annotated from these regions. The accuracies of three genomic methods, genomic BLUP (GBLUP), weighted GBLUP (wGBLUP), and BayesCπ, showed ascending trends with the number of SNPs varied from 2 k to 100 k and reach a plateau when 30 k or more SNPs were used for prediction. In addition, there was no obvious difference in predictive accuracy between the three methods when 30 k SNPs were used for prediction. In conclusion, survival against V. harveyi in leopard coral grouper belongs to a low-to-moderate-heritability trait and has a polygenic genetic architecture. Although it is feasible to apply genomic methods using the reference group constructed from nonfamily materials, a full investigation of the relationship between candidates and the reference group is necessary in leopard coral grouper commercial breeding programs, which would help maximize the predictive accuracy.