Coalitional Games and the Relevance of Gene Expression in Autism Spectrum Disorder

Background:  With the aim to provide a more powerful method for valuable signal detection in gene expression microarray experiments, a mathematical analysis of the gene expression data based into the coalitional (or co-operative) game area of game theory was recently proposed. The main advantage of this approach is the computed numerical index, called the Shapley value, which represents the relevance of each gene under a certain condition while simultaneous accounting for the expression behaviors of the other genes under the same condition, a method that, in combination with statistics, has been demonstrated to be useful for differential gene expression data analysis. 

Objectives:  To apply the coalitional game theory approach to determine the gene relevance on an ASD microarray experiment.

Methods:  Using the coalitional game method previously described by others (Moretti et al. 2008; BMC Bioinf 9:361), we have analyzed a subset from experiment GSE6575 downloaded from Gene Expression Omnibus (GEO). This subset consisted of 17 samples of Autistic patients without regression and 12 healthy children from the general population. Statistical analysis was performed with Bioconductor (http://www.bioconductor.org/) in R (http://www.r-project.org/). We detected enrichment of biological function and overrepresentation of disease(s) among gene terms using Ingenuity Pathways Analysis (IPA; http://www.ingenuity.com/).

Results:  Only two genes could be detected using standard statistical approaches. However, a relevant combination of 78 genes was obtained using the microarray game approach. IPA analysis identified 19 genes, in our set of 78 candidates, as significantly overrepresented in the general category of Neurological Disease.

Conclusions:  Our results showed that coalitional games significantly increased the power to identify candidates and that groups of these genes were associated with biological functions and disorders previously shown to be related to ASD.