A powerful Regression Module can be added to HelixTree to test allelic and haplotypic associations in the presence of confounding phenotypic variables. The Regression Module supports both linear and logistic regression. A typical workflow is to use stepwise regression to find confounding phenotypic variables, fix those regressors and then do a search for significantly associated haplotypes or individual SNPs. This regression approach is particularly powerful for overcoming the difficult challenges of population stratification. Permutation testing and adding interaction terms increase the flexibility of the analysis.

Pictured below is the regression window with parameters set to perform stepwise linear regression with predictor variables that include a three marker moving window of haplotypes, several non-genetic covariates and interactions of variables.

For more information see the following sections in the HelixTree Manual:

• Performing Regression Analysis
• Stepwise Regression
• Results from Linear Regression
• Results from Logistic Regression