Background: De novo drug discovery is a time-consuming and expensive process. Nowadays, drug repositioning is\nutilized as a common strategy to discover a new drug indication for existing drugs. This strategy is mostly used in\ncases with a limited number of candidate pairs of drugs and diseases. In other words, they are not scalable to a large\nnumber of drugs and diseases. Most of the in-silico methods mainly focus on linear approaches while non-linear\nmodels are still scarce for new indication predictions. Therefore, applying non-linear computational approaches can\noffer an opportunity to predict possible drug repositioning candidates.\nResults: In this study, we present a non-linear method for drug repositioning. We extract four drug features and two\ndisease features to find the semantic relations between drugs and diseases. We utilize deep learning to extract an\nefficient representation for each feature. These representations reduce the dimension and heterogeneity of biological\ndata. Then, we assess the performance of different combinations of drug features to introduce a pipeline for drug\nrepositioning. In the available database, there are different numbers of known drug-disease associations corresponding\nto each combination of drug features. Our assessment shows that as the numbers of drug features increase, the\nnumbers of available drugs decrease. Thus, the proposed method with large numbers of drug features is as accurate as\nsmall numbers.\nConclusion: Our pipeline predicts new indications for existing drugs systematically, in a more cost-effective way and\nshorter timeline. We assess the pipeline to discover the potential drug-disease associations based on cross-validation\nexperiments and some clinical trial studies.
Loading....