Predicted Drug-Drug Interactions in Diabetes Mellitus with Graph Convolutional Autoencoder

Rashini Maduka, Rupika Wijesinghe, Ruvan Weerasinghe

Abstract: Drug-drug interactions (DDIs) can happen when two or more drugs are taken together. Today DDIs have become a serious health issue due to adverse drug effects. In vivo and in vitro methods for identifying DDIs are time consuming and costly. Therefore, in silico based approaches are more popular in DDI identification. Most machine learning models of DDI prediction use chemical and biological properties of drugs as features. However, some properties of drugs are not available and costly to extract hence automatic feature engineering of drugs is important. Furthermore, people who suffer from diabetes mostly suffer from some other diseases as well and take more than one medicine together leading to adverse drug effects in diabetic patients. In this study we present a model with a graph convolutional auto encoder and a graph decoder using a dataset from DrugBank version 5.1.3 to predict DDIs. Main objective of the model is to identify unknown interactions between antidiabetic drugs and the drugs taken by diabetic patients for other diseases. We considered automatic feature engineering and used Known DDIs only as the input for the model. Our model has achieved 0.86 in AUC and 0.86 in AP.