Decomposition of Gene Regulatory Networks into Functional Paths and Their Matching with Microarray Gene Expression Profiles

Gene Regulatory Networks (GRNs) and DNA Microarrays (MAs) present two of the most prominent and heavily researched concepts in contemporary molecular biology and bioinformatics. GRNs model the interfering relations among gene products during the regulation of the cell function. MAs measure the simultaneous expression profile of thousands of genes. In an effort to combine these two sources of biological information, concurrent studies try either to 'build from scratch' or, focus on the expression profile of genes regulated in a specific GRN. Our work aims to discover the functional parts of GRNs by identifying consistencies and inconsistencies with respective MA gene-expression data. Specifically, GRNs are decomposed into all possible functional paths. Functional GRN paths from the repository are then matched against microarray gene-expression profiles of samples (with different clinical phenotype). In this way the underlying mechanism – the highly matched and phenotype-differentiating functional paths that guide to specific gene-expression profiles are uncovered. The discovery guides the finer re-classification of samples – with profound diagnostic and prognostic potential, and/or triggers further molecular biology research. Preliminary results on applying our methodology on a real-world microarray study and targeting the ‘Apoptosis’ gene-regulatory network are encouraging and demonstrate the suitability, efficiency and reliability of the approach.