Disease gene prediction for molecularly uncharacterized diseases

Network medicine approaches have been largely successful at increasing our knowledge of molecularly characterized diseases. Given a set of disease genes associated with a disease, neighbourhood-based methods and random walkers exploit the interactome allowing the prediction of further genes for that disease. In general, however, diseases with no known molecular basis constitute a challenge. Here we present a novel network approach to prioritize gene-disease associations that is able to also predict genes for diseases with no known molecular basis. Our method, which we have called Cardigan (ChARting DIsease Gene AssociatioNs), uses semi-supervised learning and exploits a measure of similarity between disease phenotypes. We evaluated its performance at predicting genes for both molecularly characterized and uncharacterized diseases in OMIM, using both weighted and binary interactomes, and compared it with state-of-the-art methods. Our tests, which use datasets collected at different points in time to replicate the dynamics of the disease gene discovery process, prove that Cardigan is able to accurately predict disease genes for molecularly uncharacterized diseases. Additionally, standard leave-one-out cross validation tests show how our approach outperforms state-of-the-art methods at predicting genes for molecularly characterized diseases by 14%-65%. Cardigan can also be used for disease module prediction, where it outperforms state-of-the-art methods by 87%-299%.Author summary: The elucidation of the genetic causes of diseases is central to understanding the mechanisms of action of a pathology and the development of treatments. Disease gene prediction methods streamline the discovery of the molecular basis for a disease by prioritizing genes for experimental validation. Although some methods use disease phenotype to aid the prioritization, the great majority use outdated static matrices which limits their disease coverage. Our approach uses an updatable disease phenotype similarity, and employs a non-linear transformation to define a prior probability distribution over the genes that mimics the distribution of disease genes in the interactome. Subsequently, a semi-supervised learning method establishes a prioritization ordering for all genes in the interactome, even for diseases with no known molecular basis. Our method can be used not only to obtain a better prioritization for disease-gene associations, but also for retrieving disease modules.