Epistasis Blog

From the Artificial Intelligence Innovation Lab at Cedars-Sinai Medical Center (www.epistasis.org)

Monday, May 24, 2010

A screening methodology based on Random Forests to improve the detection of gene-gene interactions

This is an interesting new paper proposing to use random forests to filter SNPs for MDR modeling of gene-gene interactions. We have seen similar results with our ReliefF-based algorithms. Removing noisy SNPs prior to MDR modeling helps cut down the number of combinations that need to be evaluated thus reducing the chances of overfitting. I am guessing RF work well on smaller numbers of SNPs but will not scale to GWAS when there are no marginal effects of the interacting loci.

De Lobel L, Geurts P, Baele G, Castro-Giner F, Kogevinas M, Van Steen K. A screening methodology based on Random Forests to improve the detection of gene-gene interactions. Eur J Hum Genet. 2010 [PubMed]

Abstract

The search for susceptibility loci in gene-gene interactions imposes a methodological and computational challenge for statisticians because of the large dimensionality inherent to the modelling of gene-gene interactions or epistasis. In an era in which genome-wide scans have become relatively common, new powerful methods are required to handle the huge amount of feasible gene-gene interactions and to weed out false positives and negatives from these results. One solution to the dimensionality problem is to reduce data by preliminary screening of markers to select the best candidates for further analysis. Ideally, this screening step is statistically independent of the testing phase. Initially developed for small numbers of markers, the Multifactor Dimensionality Reduction (MDR) method is a nonparametric, model-free data reduction technique to associate sets of markers with optimal predictive properties to disease. In this study, we examine the power of MDR in larger data sets and compare it with other approaches that are able to identify gene-gene interactions. Under various interaction models (purely and not purely epistatic), we use a Random Forest (RF)-based prescreening method, before executing MDR, to improve its performance. We find that the power of MDR increases when noisy SNPs are first removed, by creating a collection of candidate markers with RFs. We validate our technique by extensive simulation studies and by application to asthma data from the European Committee of Respiratory Health Study II.

0 Comments:

Post a Comment

<< Home