Mutation Profile Analysis of HIV Genetic Heterogeneity

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): Clinical trials have demonstrated the ability of combination therapy to curtail the amount of HIV virus (viral load) in plasma. Ironically, the use of such therapy may promote proliferation of viral drug resistant genotypes, leaving patients with few options for subsequent treatment. Alternative, genotypic-based strategies that make use of the relationship between HIV genetic alterations with some phenotypic treatment response like viral drug susceptibility are being researched to guide treatment choice. This relationship is difficult to study, due in part to its high-dimension, the quasi-species nature of HIV, and the presence of other relationships often imbedded in it, which if neglected, may lead to incorrect interpretations and treatment. For example, because of a known viral load (or other covariate) effect on observed HIV genotypes and the potential for viral load to affect the phenotype under consideration, its relationship with HIV genetic heterogeneity is likely to be confounded by viral load heterogeneity among phenotypes. The primary hypothesis to be examined is that sources of heterogeneity other than HIV genetic can be identified between phenotypic groups and their contributions to genetic associations can be estimated. New methods are proposed for extrapolating HIV genetic from other heterogeneity sources associated with phenotype, based on a non-parametric (distribution-free) approach that uses a composite (sequence pair) measure of genetic distance. In comparison to other approaches, the proposed methods are more robust, since empirical, rather than analytically modeled distributions are compared and covariate-adjusted. The methods will be developed based on a profile analysis-type setting in which a series of related hypotheses for examining the relationship of interest are constructed. The specific aims are: (1) to construct a measure-theoretic framework for modeling quasi-species heterogeneity associated with covariates using distribution-free approaches, including distance-based; (2) to develop distribution-free approaches for identifying covariates that alter the relationship between HIV genetic heterogeneity and phenotype and to adjust this relationship for such covariates; (3) to extend the approaches in Aims 1 and 2 to address multivariate analysis of gene regions and temporal trends in HIV genetic heterogeneity associated with phenotype, in the presence of covariates that depend upon these continuums; (4) to characterize HIV genetic heterogeneity associated with phenotype by developing methods that translate composite genetic distance measures into region, time and location effects; (5) to assess the sensitivity of the methods developed in Aims 1-3 to the type of phenotypic populations with respect to their genetic compositions; (6) to computerize the methods in Aims 1-4 into a free, internet-accessible software; and (7) to communicate statistical tests to clinicians/virologists using paradigms that translate their meaning into genetic settings.