From Genotype to Phenotype从基因型到临床表现型.ppt

* * * * P-values from trend test are plotted against p-values from a test which corrects for population structure. SNPs at which signals of association is affected by correction for population structure appear as points off the diagonal. Trend test based on logistic regression w/ or w/o covariates (First two PCs of geographic data). * Roughly mirrors actual geography of UK Signfiicant overlap in 1st two principal components * * * * * * * * * Overview of impact of genetics from single mutation to entire genome * * * * * * * * * Age, gender, physiological pathological conditions, intake of food an drugs, and physical activity * * * * * * * * Overview of impact of genetics from single mutation to entire genome * * * * * * * * * * * * * * * * * * * * * * Step 2 of the Analysis: Linking Genotype to Phenotype Bayesian Regression Model Regress on genotypes to arrive at estimates of phenotype Genotype model is additive 2 questions to answer Are any of the non-zero? Which are non-zero and how big are they? Simulations show that imputation method can be just as powerful as if all SNPS were sequenced Genotype to Phenotype Motivation Background Genetics of Cardiovascular Disease Genome-Wide Association Studies Future Directions Genotypic biomarkers still can’t substitute for Phenotypic biomarkers Phenotypic biomarkers such as cholesterol, blood glucose, pressure, etc. integrate large number of genetic and nongenetic influences Single genetic polymorphisms not useful in diagnosing complex diseases Need to integrate multiple polymorphic signals to be clinically useful Need to understand disease at systems-level GWA immediately useful in establishing causality of phenotypic markers Association of genetic variants with both levels of phenotypic biomarker and disease risk is an argument for causality For example, circulating levels of inflammatory biomarkers such as various cytokines are known to be increase in the course of atherosclerosis Could be u