| In genome-wide association analysis(GWAS),survival traits are generally described by binary survival and survival time.Because both phenotypes do not follow the normal distribution,the traditional linear mixed-model association analysis is not suitable for identifying quantitative trait nucleotides(QTNs).Therefore,the researchers embedded the parametric and semi-parametric models in survival analysis into the framework of interval location of quantitative traits to locate survival traits.Among,the Cox proportional hazards model is a vital model for analyzing survival traits.Due to the lack of effective methods for analyzing survival traits of aquatic animals,this paper applied the Cox proportional risk model in survival analysis to the genomewide association analysis of aquatic animal genetics and proposed a new efficient genome-wide association analysis method for survival traits.Based on the Cox mixed effect models(CMEMs),this method combines the genome-wide rapid association using the mixed model and regression(GRAMMAR)method with survival traits,and the estimated genome breeding values(GBVs)are regarded as a known predictor.Then,by using genome control to correct the association test statistics,GRAMMAR-Lambda is successfully extended to the field of survival traits.Compared with other survival analysis methods,this method improves the statistical effectiveness of the correlation test and greatly improves the computational efficiency of the correlation test of survival traits.This study simulated survival traits under different combinations of censoring rates and quantitative trait nucleotides based on the human dataset.Computer simulation experiments show that:(1)When estimating GBVs,genomic heritability does not need to be estimated in advance and can be given within a certain range.Generally,it can be set to 0.5 by default,or it can be set to empirical heritability for testing traits.(2)When calculating the kinship matrix,GBVs can be estimated by using a small number of sampling markers,which can achieve the same effect as full markers.(3)Nearly perfect genomic control,with the same statistical efficacy as the Cox Mixed-effects Models for Genome-wide Association Studies(COXMEG)method,the computing speed is dozens of times faster than COXMEG.Genome-wide association analysis of traits between rainbow trout(Oncorhynchus mykiss)rickettsia disease,human actual disease(Type 1 diabetes,Malignant neoplasm of prostate,Chronic ischemic heart,and Stroke),and rainbow trout infectious pancreatic necrosis confirmed that GRAMMAR-Lambda has better statistical power and computational speed than COXMEG method in populations with higher complex structure faster;In groups with low complex structure,GRAMMAR-Lambda has the same statistical power as the COXMEG method and is tens to hundreds of times faster.This provides a new method basis for GWAS research and genetic structure analysis of survival traits of other aquatic animals and has important biological significance. |
