The main theme of this research is on haplotype, multilocus and general genetic association methods. We have been developing haplotype association mapping approach that relax population assumptions, such as random mating, and bypass estimation of the haplotype phase. Compared to competitors, the approach gains substantial power under low linkage disequilibrium and when there is non-independence of haplotypic effects. The approach is based on estimation of joint frequencies of alleles at multiple genetic loci. The difference between traditional approaches and this approach lies in that the traditional approaches consider joint frequencies of alleles that reside on the same haplotype, whereas we consider the sum of joint frequencies that reside on the same as well as on two different haplotypes within individuals. This allows to capture direct haplotypic effects as well as joint effects of alleles that reside on two different haplotypes within an individuals, which reflect the diplotypic effects. The approach is applicable for testing interactions between distant genetic loci. For example, it can be applied to sets of SNPs that span separate haplotype blocks, and it is also useful for detecting interactions between genetic loci that reside on different chromosomes. While the method is powerful in detecting direct haplotypic effects, it is also capable of capturing epistatic effects due to interactions between two haplotypes in a diplotype. Covariates, such as environmental exposures, can be readily incorporated with this approach. The main development has been around the case-control and random population sampling designs. There are straightforward extensions to family-based association mapping designs. These extensions make the approach robust against confounding due to population stratification. We have applied the method in study concerned with discovery of novel functional variants in the mu-opioid receptor, which is the principal receptor target for both endogenous and exogenous opioid analgesics. There are substantial individual differences in human responses to painful stimuli and to opiate drugs that are attributed to genetic variations in the mu-opioid receptor. Opioid drugs are widely used to treat both acute and chronic pain, but the benefits are undermined due to individual variation in side effects and the degree of response. Our approach helped to identify novel genetic variants that offer insight into individual responses to morphine. This may lead to development of a new class of opioids and to diagnostic tests for prediction of individual risk for inadequate or adverse response. In an ongoing collaboration with Dr. Honglei Chen this method is being applied to Parkinson's disease genome scan data. Other ongoing research include development of statistical approaches to address multiplicity issues in whole genome scans. This research includes investigation of novel approaches aimed to improve ranks of true positives in whole genome scans (in collaboration with Dr. Jack Taylor), and development of new tests for the shared control design, where a common control group of individuals is contrasted against several independent case groups that may have the same, different, or related diseases.
Showing the most recent 10 out of 29 publications
