Overview of approach to identify novel drug-disease (Dr-D) associations.

Gene-disease associations from 10 sources are first integrated and ranked. These scored associations are then integrated with protein, gene, disease and drug data to give an integrated dataset. A therapeutic area of application is then identified before the dataset is mined for instances of a semantic subgraph whose mappings contain inferred dd associations. Finally, any dd associations that are likely side effects (SEs) are filtered using the MeSH distance measure, before all ranked dd associations are returned.