For Manuscript Submission, Check or Review Login please go to Submission Websites List.
For the academic login, please select your country in the dropdown list. You will be redirected to verify your credentials.
Power of Data Mining Methods to Detect Genetic Associations and InteractionsMolinaro A.M.a · Carriero N.b · Bjornson R.b · Hartge P.c · Rothman N.c · Chatterjee N.c
aDivision of Biostatistics, School of Public Health, and bDepartment of Computer Science, Yale University, New Haven, Conn., and cDivision of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of Health, Rockville, Md., USA Corresponding Author
Annette M. Molinaro
Division of Biostatistics
School of Public Health, Yale University
New Haven, CT 06519 (USA)
Background: Genetic association studies, thus far, have focused on the analysis of individual main effects of SNP markers. Nonetheless, there is a clear need for modeling epistasis or gene-gene interactions to better understand the biologic basis of existing associations. Tree-based methods have been widely studied as tools for building prediction models based on complex variable interactions. An understanding of the power of such methods for the discovery of genetic associations in the presence of complex interactions is of great importance. Here, we systematically evaluate the power of three leading algorithms: random forests (RF), Monte Carlo logic regression (MCLR), and multifactor dimensionality reduction (MDR). Methods: We use the algorithm-specific variable importance measures (VIMs) as statistics and employ permutation-based resampling to generate the null distribution and associated p values. The power of the three is assessed via simulation studies. Additionally, in a data analysis, we evaluate the associations between individual SNPs in pro-inflammatory and immunoregulatory genes and the risk of non-Hodgkin lymphoma. Results: The power of RF is highest in all simulation models, that of MCLR is similar to RF in half, and that of MDR is consistently the lowest. Conclusions: Our study indicates that the power of RF VIMs is most reliable. However, in addition to tuning parameters, the power of RF is notably influenced by the type of variable (continuous vs. categorical) and the chosen VIM.
© 2011 S. Karger AG, Basel