Haplotype-based Classifiers to Predict Individual Susceptibility to Complex Diseases - An Example for Multiple Sclerosis

The enormous amount of genetic data that is currently being produced with the explosion of genome-wide association studies is yielding an important effort in the construction of genetic-based predictive models for individual susceptibility to complex diseases. However, a constant pattern of low accuracy is observed in most of them. We hypothesize that a main cause of their low accuracy is the strong reduction of genetic information considered by the classifiers, and propose a three-fold solution that considers haplotype instead of genotype individual data, whole-genome markers instead of a more stringent selection and several-marker risk variants instead of only one or two. We have compared the performance of our approach with current approaches to predict individual genetic risk to multiple sclerosis, and have found that our method yielded significantly more accurate classifiers.