Gene-disease prioritization through biomedical networks

Identifying associations between genes and diseases has been an important research challenge in biomedicine, since this will help to create new diagnostics and treatments. This paper describes a computational method that ranks a list of genes according to their level of association to a given disease. In a first step, it constructs a graph representation of biomedical terms such as genes, pathways, homologies, ontologies, gene expression data and literature. Then the implemented algorithm runs over this network taking into consideration relations between the biomedical terms. The results obtained confirm the advantages of the proposed method when used to prioritize genes from a gene expression study. The solution presented here represents major progress when compared with previous methods, hopefully allowing faster and easier establishment of gene-disease associations.