Estimating Realized Heritability for Growth in Zhikong Scallop (chlamys Farreri) Using Genome-Wide Complex Trait Analysis

Haobing Guo,Qifan Zeng,Yangping Li,Yangfan Wang,Zhihui Chen,Ping Lin,Shi Wang,Zhenmin Bao

DOI: https://doi.org/10.1016/j.aquaculture.2018.07.046

IF: 5.135

2018-01-01

Aquaculture

Abstract:In selective breeding, a central parameter in summarizing the proportion of variance due to genetics for the purpose of predicting gains from selection is realized heritability (h(2)). We applied the genome-wide complex trait analysis (GCTA) method to genome-wide SNP data obtained by 2b-RAD reduced-representation genotyping and phenotypic data for four traits (shell length, shell height, shell width and whole wet weight) with the aim of establishing the heritability for growth in Zhikong scallop (Chlamys farreri) in a selective breeding program. The GCTA-based heritabilities of 0.42 (S.E. 0.09) for shell length, 0.47 (S.E. 0.07) for shell height, 0.54 (S.E. 0.11) for shell width and 0.28 (S.E. 0.03) for whole wet weight, which were estimated with uncommon SNPs (26,471 SNPs with MAF > 2%), were close to and strongly correlated (r = 0.957) with traditional estimates of realized heritability and in the moderate-to-high range, in line with values previously obtained for growth rates in bivalves. h(GCTA)2 with 20,000 and 10,000 SNPs were very close to estimates with uncommon SNPs and appeared relatively robust to SNP number. Removing causal SNPs had little effect on obtaining reliable estimates of h(GCTA)2, suggesting that causal SNPs is not necessary for accurate estimates of h(2). Chromosome-wise heritability estimates suggested that the genetic contribution to growth complex traits is scattered across the genome and driven by many loci with small effect rather than a few causal loci with a large effect. These results indicate that the genome-wide complex trait analysis method may be useful for estimating realized heritability for growth in Zhikong scallop with SNPs obtained by reduced-representation genotyping approaches, which is less expensive and faster than full-genome sequencing for non-model species growing in natural environments.