Aims 1 and 2? ? Theoretical work during the past year focused on linear regression based tests of associations for quantitative traits, alternatives to theoretically based determination of type I error rates and the multiple testing issue. Work on the revised Regression of Offspring on Mid-Parent (ROMP_rev) method has been completed. Like the original ROMP method, ROMP_rev is a hybrid of the traditional regression of offspring on mid-parent model used to estimate heritability and the ANOVA model used to test for association. ROMP_rev was developed within the context of collapsibility in linear regression models in order to determine the standard error of the difference between two regressions. The ROMP_rev method provides a test of association between a quantitative trait and a marker locus and also provides an estimate of the locus-specific heritability. Unlike the TDT for qualitative traits, ROMP_rev requires phenotype data on parents and offspring, but genotype data for the offspring only. This method was developed for studies where phenotype, but not genotype, data are available for the parents Roy-Gagnon et al., Ann Hum Genet, in press. ? ? With respect to type I error rates, simulation studies have been performed during the past year to determine whether extremely small p-values, such as those seen in Genome-Wide-Association (GWA) studies, are due to random errors or due to deviations from the assumptions underlying the models. Although a number of investigators believe that extremely small p-values reflect either the failure of normal distribution theory or due to failures of assumptions such as non-Hardy-Weinberg equilibrium, results from the simulations suggest that of the models considered, tests of association are extremely robust to deviations from the assumptions underlying the model, and that the distribution of extremely small type I errors is what is expected.? ? Theoretical work has also begun on methods that use multiple regression and regions defined by the HapMap to address the non-independence of SNP data in linkage disequilibrium.? ? HapMap Hotspot sequences? ? With the identification of a large set of hotspots in the human genome, the frequency and distribution of specific sequence motifs can be compared in hotspot and cold spot regions. To date, relatively few sequence motifs have been associated with hotspots. The frequency of every 7 bp motif was compared in hotspot relative to coldspot regions, and paired t-tests were calculated; however, based on previous unpublished work, sequence composition was considered at the purine/pyrimidine level (RY) rather than at the ACGT nucleotide level. Of all possible 7-mers at the purine/pyrimidine level, 0.86 fewer copies of the RYYRRYR/YRYYRRY motifs were found in the hotspot relative to the coldspot regions (p < 1.37E-23), while the RRRRRRR/YYYYYYY motifs were more frequent in the hotspot than in the coldspot regions (1.82 copies, p<1.83E-30). When the regions flanking hotspots were compared to the regions flanking the coldspots, the differences between the frequencies of both motifs in hotspots relative to coldspots rapidly approached zero as the distance from the hotspot boundary increased. Permutation tests, where motif frequencies in hotspot or coldspot regions were compared to those of size-matched random sequence from the genome, confirm that the RYYRRYR/YRYYRRY motifs were less frequent in hotspots while the RRRRRRR/YYYYYYY motifs were more frequent in coldspots. ? ? Aim 3? ? Genotyping for several ongoing collaborations continues. These include:? 1) fine mapping (SNPs) and determination of copy number polymorphisms in candidate regions for the familial scoliosis project? 2) fine mapping for candidate regions for the STAR-D and NIMH collaboration with Drs. John Rush and Francis McMahon on genes for response to citalipram? 3) fine mapping and SRTP and SNP panels for tests of linkage and association in the JHU GeneSTAR / NHLBI PROGENI gene-environment study. ? 4) candidate gene studies and linkage panels in the Madras diabetes study.? ? Much of the effort during the past year focused on statistical genetic analysis in collaborative studies.? ? Familial idiopathic scoliosis. Several analyses focusing on candidate regions and phenotypic subsets have been completed and manuscripts have either been submitted or are in preparation. These include:? 1) Genetic analysis of patients with triple curves. In this study of susceptibility loci in FIS families with at least one individual with a triple curve, candidate regions were identified on chromosomes 6 (27 Mb) and 10 (9 Mb).? 2) FIS replication study. Statistical genetic analysis of a replication sample of families with familial idiopathic scoliosis with characteristics nearly identical to those of the sample analyzed in Miller et al. 2005 suggest that of the 11 regions identified in the original study, at least 7 of these regions can be considered to be true replications Behnemann et al. in preparation. The replication of the original candidate regions is another critical step in identifying candidate loci in subsequent association studies. ? ? Other ongoing analyses from the familial idiopathic scoliosis project included analyses of candidate regions on chromosomes 9 and 16, and a subset based on the presence of males with surgery.? ? ? Genetic analysis of drug treatment response in the Sequenced Treatment Alternatives for Depression (STAR*D). ? ? As part of collaborations with investigators from the Southwestern Medical Center (John Rush) and NIMH (Francis McMahon), statistical analyses have focused on several candidate genes. In addition to the 5HTR2A receptor reported in McMahon et al. Am J Hum Genet, 2006 other associations reported include antidepressant treatment and the GRIK4 locus Paddock et al. Am J Psychiatry, 2007, and genetic markers associated with suicidal ideation Laje et al. Am J Psychiatry, 2007.? ? GenStar? ? Genotyping data from the JHU/NHLBI PROGENI GenSTAR project became available in late 2005. Statistical genetic analyses are underway for this very large collaboration. Tests of both linkage and association are being performed in these families to identify genetic effects and gene-environment effects due to an aspirin intervention. Two projects focusing on estimating heritabilities of variables related to inflammation and thrombosis have been completed. The first focused on estimating the heritability of platelet response to aspirin and cyclooxygenase-1 Faraday et al. Circulation 2007. The second reported heritabilities for variables related to platelet function Bary et al. Thromb Haemost 2007. ? ? Genetics of type II diabetes in India. The collection of family data, phenotypes and blood for genotyping is nearly completed. The goal of the project is to recruit 30 large multiplex families to study the genetics of type II diabetes in an Indian population. Approximately 24 families have been recruited at this point. DIR funds were used to continue a contract with collaborators at the Madras Diabetes Research Foundation for family recruitment, and statistical analysis for tests of association of SNPs in candidate genes were performed.
Showing the most recent 10 out of 23 publications
