Multilocus statistics to uncover epistasis and heterogeneity in complex diseases: revisiting a set of multiple sclerosis data

Abstract

New statistics are developed to gather the contribution of many alleles at different loci to common diseases. Both inferential and descriptive statistics are included in order to uncover epistatic effects as well as heterogeneity. The problem of multiple testing is circumvented by considering a global null hypothesis. Global testing is supplemented by descriptive methods that make use of measures like odds ratio or the Pvalue of individually tested allele combinations. Visualization helps to reflect complex data sets. The methods described here have been scrutinized by statistical simulations, and we show that power gains can be substantial as compared to single locus statistics. Typing data of multiple sclerosis patients and controls are investigated, representing an example of larger scale information in screening candidate genes for their impact on complex diseases. New insights emerge from this data set demonstrating genetic heterogeneity and evidence for epistasis.

Efficient Algorithms for Detecting Genetic Interactions in Genome-Wide Association Study

“…The upper bounds used in these methods do not work for other statistical tests, such as chi-square test, G-test, information-theoretic association measurements, and trend test (Balding (2006); Pagano and Gauvreau (2000); Thomas (2004)), which are also routinely used by researchers. In addition, new statistics for epistasis detection are continually emerging in the literature (Bohringer et al (2003); Dong and et al (2008); Zhao et al (2005)). Therefore, it is desirable to develop a general model that supports a variety of statistical tests.…”