Genomic inbreeding evaluation and its effects on phenotype and selection accuracy of resistance to Edwardsiella tarda in Turbot (Scophthalmus maximus)

Abstract

To accurately evaluate the effects of inbreeding on phenotypes and selection accuracy, this study conducted resequencing on a population of turbot (Scophthalmus maximus) involved in resistance testing against Edwardsiella tarda and calculated individual inbreeding coefficients (F) based on SNP homozygosity ($F_{\mathrm{HOM}}$), diagonal of the SNP-based genomic relationship matrix ($F_{\mathrm{GRM}}$), and runs of homozygosity (ROH) with varying length thresholds (10 kb, 100 kb, 500 kb, 1000 kb; denoted as $F_{\mathrm{ROH}10}$, $F_{\mathrm{ROH}100}$, $F_{\mathrm{ROH}500}$, $F_{\mathrm{ROH}1000}$). The impact of F derived from these methods on phenotypic traits was assessed, and the potential improvement in genomic selection accuracy by incorporating inbreeding depression (ID) effects into the genetic evaluation model was examined. The average values of $F_{\mathrm{HOM}}$ and $F_{\mathrm{GRM}}$ were near zero (−0.006 ± 0.107 and 0.004 ± 0.064, respectively), while $F_{\mathrm{ROH}10}$ exhibited the highest mean value (0.103 ± 0.064). As the ROH fragment length increased, the ROH-based F gradually decreased, with $F_{\mathrm{ROH}1000}$ averaging 0.005 ± 0.011. Regression analysis revealed negative regression coefficients across all methods, but only $F_{\mathrm{ROH}100}$ and $F_{\mathrm{ROH}500}$ showed significant effects on phenotypic traits (p < 0.05 and p < 0.01, respectively). $F_{\mathrm{ROH}1000}$ exhibited a marginal association with the phenotype (p < 0.10). The estimated ID effects on resistance to E. tarda were largest for $F_{\mathrm{ROH}1000}$ (10.871 %), followed by $F_{\mathrm{ROH}500}$ (6.033 %) and $F_{\mathrm{ROH}100}$ (2.443 %). The prediction accuracy of GBLUP was 0.539. The accuracies of GBLUP-$F_{\mathrm{ROH}100}$ and GBLUP-$F_{\mathrm{ROH}1000}$ were nearly identical to that of GBLUP (0.538 and 0.539, respectively), while GBLUP-$F_{\mathrm{ROH}500}$ showed a slight increase (0.543), though non-significant. These results indicate that while genomic inbreeding (especially $F_{\mathrm{ROH}}$) negatively impacts traits, the interference of inbreeding effects on selection accuracy resistance to E. tarda can be kept at minimal levels.