An application of Random Forests to a genome-wide association dataset: Methodological considerations & new findings

This study presents one of the first illustrations of successfully analyzing GWA data with a machine learning algorithm. It is shown that RF is computationally feasible for GWA data and the results obtained make biologic sense based on previous studies. More importantly, new genes were identified as potentially being associated with MS, suggesting new avenues of investigation for this complex disease.