Drug repurposing and prediction of multiple interaction types via graph embedding

Abstract

Background: Finding drugs that can interact with a specific target to induce a desired therapeutic outcome is key deliverable in drug discovery for targeted treatment. Therefore, both identifying new drug-target links, as well as delineating the type of drug interaction, are important in drug repurposing studies.

Results: A computational drug repurposing approach was proposed to predict novel drug-target interactions (DTIs), as well as to predict the type of interaction induced. The methodology is based on mining a heterogeneous graph that integrates drug-drug and protein-protein similarity networks, together with verified drug-disease and protein-disease associations. In order to extract appropriate features, the three-layer heterogeneous graph was mapped to low dimensional vectors using node embedding principles. The DTI prediction problem was formulated as a multi-label, multi-class classification task, aiming to determine drug modes of action. DTIs were defined by concatenating pairs of drug and target vectors extracted from graph embedding, which were used as input to classification via gradient boosted trees, where a model is trained to predict the type of interaction. After validating the prediction ability of DT2Vec+, a comprehensive analysis of all unknown DTIs was conducted to predict the degree and type of interaction. Finally, the model was applied to propose potential approved drugs to target cancer-specific biomarkers.

Conclusion: DT2Vec+ showed promising results in predicting type of DTI, which was achieved via integrating and mapping triplet drug-target-disease association graphs into low-dimensional dense vectors. To our knowledge, this is the first approach that addresses prediction between drugs and targets across six interaction types.

Keywords: Drug discovery; Drug repurposing; Drug-target interaction; Machine learning; Network embedding.

Fig. 1

DT2Vec+ pipeline. (a-1,2,3) integrating drug–drug (DDS) and protein–protein (PPS) similarity graphs with drug-disease (DDis) and disease-protein (DisP) association graph as input of embedding method to low dimensional vectors. (a-4) Drug–target interaction graph with different edge types. b Graph-embedding developed by node2vec to map nodes to vectors (in this figure, nodes are shown mapped to 2D-vector, x and y). c Known drug–target interactions (six types of interactions) were divided into 10% independent dataset (external testset) and 90% internal test and train (tenfold cross-validation). d Drug and protein vectors were concatenated and labelled using one-hot encoding and an XGBoost model was trained on each label using cross-validation. The best model over the tenfold cross-validation on the internal testset was selected and applied on the external testset. The XGBoost model in c, d was repeated 5 times and the average performance of internal and external testsets was reported

Fig. 2

Drugs, diseases, proteins and DTI visualisation. a PCA of drugs, diseases, and proteins vectors extracted from heterogeneous association graph. b PCA of DTIs mapped vectors by concatenating drug–target vectors

Fig. 3

Interaction type of all drug–target pairs. The heatmap shows the mapping of known DTI interactions (red and purple) and predicted interactions (blue and lilac). Each interaction can have multiple types of interactions, which were coloured darker

Fig. 4

Top novel drug–target interactions. Top 20 DTIs predicted using DT2Vec++ for each type of interaction coloured based on the type of interaction