In Fall 2012 I am teaching two classes, EPI 215 (Fall-1) and EPI 507 (Fall-2).
EPI 215 (Advanced Topics in Case-Control and Cohort Studies), co-taught with Dr. Lori Chibnik, extends the parametric regression models covered in EPI 204. Topics include techniques for modeling continuous and polytomous exposures, methods to account for missing data, doubly-robust modeling, and high dimensional data analysis, overfitting and methods to assess or avoid it, risk prediction, and sample size calculations. Emphasis is on applications with limited introduction to theory that underlies these techniques. Familiarity with SAS is desirable; basis R programming will be introduced in lab.
EPI 507 (Genetic Epidemiology), co-taught with Dr. Liming Liang, introduces the basic principles and methods of genetic epidemiology. After a brief review of the history of genetic epidemiology, methods for the study of both high penetrance and low penetrance alleles will be described and discussed. Methods of analysis of genome-wide association studies are a particular focus. Examples of the contribution of genetic analysis to major diseases will be reviewed.
Previously taught courses at HSPH include:
BIO 227: Fundamental Concepts of Gene Mapping (2004-2005)
Grading Instructor (Co-Instructor Christoph Lange)
This course introduced students to the diverse statistical methods used throughout the process of genetic epidemiology, from familial aggregation and segregation studies to linkage scans candidate-gene association studies. Topics covered included multipoint and model-free linkage analysis, linkage disequilibrium, family-based and population-based association test, and study design. Ongoing research into the genetics of asthma and cancer were used to illustrate basic principles. Homework included analysis projects to familiarize students with state-of-the-art software for linkage analysis, family-based association tests, and case-control studies. The goal of this class was for students to leave with a basic understanding of how to read and evaluate statistical studies of genetics epidemiology.
BIO 228: Statistical Genetics of Complex Human Disease (2005-2006)
Co-Instructor with Christoph Lange
This course concentrated on the design and analysis of complex-disease association studies. It started with a review of key concepts in genetic epidemiology (population genetics, population stratification, linkage disequilibrium, “tagging” SNPs) and then proceeded in two parts: population-based studies and family-based studies. Each part covered problems of design and analysis, among them: choice of test statistic, inferring and analyzing phased haplotypes from unphased genotypes, multiple comparisons, and gene-environment interactions. Homework consisted of two analysis projects (one population-based and one family-based) designed to give students hands-on experience with current applications and software.
EPI 292: Advanced Topics in Epidemiologic Methods (2006)
This class provided an in-depth investigation of statistical methods for drawing causal inferences from observational studies. Informal epidemiologic concepts such as confounding, selection bias, overall effects, direct effects, and intermediate variables were formally defined within the context of a counterfactual causal model and with the help of causal diagrams. Methods for the analysis of the causal effects of time-varying exposures in the presence of time dependent covariates that are simultaneously confounders and intermediate variables were emphasized. These methods included g-computation algorithm estimators, inverse probability weighted estimators of marginal structural models, g-estimation of structural nested models. As a practicum, students reanalyzed data sets using the above methods.
EPI 293: Analysis of Genetic Association Studies (2006-2008, 2010)
This course introduced the conceptual and practical tools needed for genetic association studies using unrelated subjects. Students gained hands-on experience with a range of analytic tools and software packages as part of a class project which gave them the opportunity to design and analyze an association study. This project required students to tackle real-world problems such as marker selection, potential multiple comparisons issues due to multiple markers and multiple outcomes, and missing data. Lectures and selected readings presented key ideas (such as linkage disequilibrium, “tagging SNPs,” haplotypes, population stratification and epistasis) and appropriate statistical methods.