Hierarchical Mixed Model For Genome-wide Association Analysis Of Animal Growth Trajectories

Random regression model(RRM)can flexibly characterize changes in genetic effects of markers and polygenes on longitudinal phenotype by Legendre polynomials,but complex computation hinders application of RRM into genome-wide association analysis for growth trajectory.To improve computing efficiency for longitudinal association analysis,we divide genomic RRM into the two hierarchies that first fits growth trajectory of each individual using least square or a random regression model,and then associates phenotypic regressions with genetic markers using a multivariate mixed model(mv LMM).At the first hierarchy,individuals' growth curves and heterogeneous residual variance function require to be chosen according to Bayesian information criterion.At the second hierarchy,multivariate mixed association analysis is canonically transformed to multiple univariate association analyses by using variance components estimated in advance,which enlarged number of phenotypic regressions.Within the framework of longitudinal association analysis,joint association analysis for quantitative trait nucleotide(QTN)candidates from multiple testing is straightforwardly made to improve statistical power to detect QTNs governing growth trajectory.For understanding genetic basis substantial difference in body weights caused by ecological adaptation and evolution,a F2 resource population has been constructed by crossing between WSB and GI strains.11833 SNP makers extracted from 1215 F2 mice by strict quality control are used to do genome-wide association analysis for growth trajectory of body weights.We fit individuals' growth trajectories with Legendre polynomials and Richards growth curves,and then associate phenotypic regressions with genetic makers.As the results,one QTN:UNC18906603 on Chromosome 10 is found by a test at once,which positively impacts on growth trajectory of body weights in the pattern of parabola during the measurement.By joint association analysis,moreover,an additional QTN:JAX00177214 is found on chromosome 20,which has negatively genetic effects with the same pattern as the QTN(UNC18906603).