Research
Direct Estimation of Effects and Tests of Their Statistical Significance for the Case of One Autosomal Locus with Two Alleles
In previous papers, see the references, the author introduced methods for estimating effects directly in samples of individuals whose genomes had be sequenced for the cases of one and two or more quantitative traits. In these papers, no attention was given to developing procedures of testing the statistical significance of the estimated effects. This paper is devoted to the development of statistical tests of significance of estimated effects for the simple case of one autosomal locus with two alleles, using Monte Carlo simulation methods. Because no real data was available to the author, artificial data for the three genotypes was simulated by using a Monte Carlo simulation procedures with fixed sample size for each genotypes as well as expectations and variances. In all cases considered, the null hypothesis was described in detail so as to inform a reader on the basic concepts underlying the proposed tests of statistical significance.
Estimating Effects and Variance Components in Models of Quantitative Genetics in an Era Sequenced Genomes
As in many other areas of research in genetics, the availability sequenced genomes in samples of individuals has revolutionized the study of quantitative traits, because researches have developed statistical evidence regarding the locations of genomic regions, loci, that have been implicated with the expression of a quantitative trait or traits. Therefore, in cases in which it is possible to develop operational definitions of at least two alleles at each locus, genomic regions, it becomes possible to identify the genotype of each individual with respect to a set of loci that have been shown in other experiments to influence the expression of a quantitative trait. As will be shown in this paper, by knowing the genotype of each individual in a sample with respect to a set of identified loci, it is now possible to directly estimate effects that are measures of not only intra-allelic interactions at each locus under consideration but also various types of epistatic effects that are measures of interactions among alleles at different loci, governing the inheritance of a quantitative trait. These straight forward methods of estimation differ from those used in classical quantitative genetics, because such effects and corresponding variance components could be estimated indirectly, using analysis of variance procedures or some version of general linear models that have been and are widely in statistical genetics. The direct methods of estimation described in this paper, show promise towards shifting the working paradigm that has been used in classical models of the genetics of quantitative traits involving the estimation of variance components to a simpler and more direct approach.
