The leopard coral grouper (Plectropomus leopardus) is a warm-water reef fish belonging to the family Serranidae, subfamily Epinephelinae, and genus Plectropomus. Vibrio harveyi is the main pathogen that causes "rot disease" in leopard coral grouper, which is a major threat to the sustainable development of its aquaculture industry. The disease is highly prevalent from June to August, severely affecting the species' aquaculture. Therefore, developing disease-resistant strains is an urgent issue to address. However, artificial breeding techniques for leopard coral grouper currently cannot establish a family lineage through one-on-one artificial insemination, which makes traditional breeding methods relying on clear pedigree difficult. Given the successful breeding cases of disease-resistant fish species with or without pedigree, genome selection breeding technology may become the key technology for cultivating disease-resistant leopard coral grouper breeding.
In genetic selection, the genetic parameters of target traits are important reference factors for specifying breeding programs. To evaluate the genetic parameters of leopard coral grouper resistance to V. harveyi, this study constructed a genome-relatedness matrix based on high-density single nucleotide polymorphisms and used four models (BLM, BTM, LLM, LTM) to fit two disease-resistant phenotypes (Test-day trait, TDS; Bivariate survival trait, TS), and used restricted maximum likelihood (REML) to estimate variance components. The results show that the genetic heritability of leopard coral grouper resistance to V. harveyi ranges from 0.182 to 0.486, which belongs to the medium to high genetic heritability range. Additive genetic variance ranges from 0.071 to 0.262. Among them, the genetic heritability estimated by the linear model is 0.382 and 0.476, while that estimated by the threshold model is 0.182 and 0.207. These results indicate that leopard coral grouper's resistance to V. harveyi can be improved through genetic breeding.
In this study, the linear model (BLM, LLM) obtained higher genetic heritability estimates and more accurate genomic estimated breeding values (GEBV) predictions than the threshold model (BTM, LTM). However, regardless of the model used, the correlation coefficient between GEBV rankings under the same phenotype definition exceeds 0.9, indicating that their impact on GEBV ranking is not significant. Compared to using cross-sectional models (BLM, BTM), a large number of leopard coral grouper GEBVs are rearranged in the longitudinal linear model (LLM) results. There is a strong correlation between the linear threshold model (LLM) and the phenotype (TS), indicating that the prediction effect of the LLM model is excellent. Therefore, when breeding leopard coral grouper for V. harveyi-resistant traits, the longitudinal linear model (LLM) should be considered.
Regardless of using linear or threshold models, it is observed in this study that using longitudinal models (LLM, LTM) to estimate genetic heritability is higher than that obtained from cross-sectional models (BLM, BTM), which may be due to death time explaining different components of fish disease resistance. In the longitudinal model, the genetic component determined by death time can be effectively used, but in the cross-sectional model, this effect is included in the residual. Similar to the genetic heritability results, the longitudinal model obtains more accurate GEBV estimates than the cross-sectional model. These results indicate that TDS may demonstrate a more accurate definition relative to TS when evaluating leopard coral grouper's resistance to V. harveyi.
In addition, compared to the threshold model, the linear model performs better in GEBV prediction in this study, and higher genetic heritability estimates are obtained. Although most previous studies on disease resistance traits have inconsistent genetic heritability estimates between the threshold model and the linear model, some studies support the above conclusions. The result may be due to differences in information processing between different models that lead to different results. In this study, the additive genetic variance obtained by the threshold model (BTM, LTM) was 0.222~0.262, and the additive genetic variance obtained by the linear model (LLM, BLM) was 0.071~0.086. It is common sense that the additive genetic variance obtained by the threshold model is higher than that obtained by the linear model. Moreover, the residual variance based on linear model fitting is small. Combined with the above results, this study believes that when threshold traits are regarded as normally distributed data and the linear model is used for fitting, the residual variance is underestimated due to the underfitting of the model, which results in a high heritability estimation of the linear model.
This study aimed to estimate the genetic parameters of leopard coral grouper resistance to V. harveyi, using infection test data of leopard coral grouper injected with V. harveyi and constructing individual genotype relationship matrix based on single-nucleotide polymorphisms. The genetic heritability of leopard coral grouper resistance to V. harveyi was estimated to be between 0.182 and 0.486, through the comparison of different models and phenotype definitions. Both the linear model (0.382 and 0.476) and the threshold model (0.182 and 0.207) were used for the estimation of genetic heritability. The estimated genetic heritability level belongs to a medium genetic heritability range. These results can be utilized for improving the target traits of leopard coral grouper, specifically its resistance to V. harveyi. Additionally, the longitudinal model provided more genetic information than the cross-sectional model in the genetic heritability evaluation of leopard coral grouper resistance to V. harveyi. This study supplements the genetic parameter estimation of leopard coral grouper resistance to V. harveyi, which provides a reference for the selection of V. harveyi-resistant leopard coral grouper breeding. |