A Computational-Based Method for Predicting Drug–Target Interactions by Using Stacked Autoencoder Deep Neural Network

Abstract

Identifying the interaction between drugs and target proteins is an important area of drug research, which provides a broad prospect for low-risk and faster drug development. However, due to the limitations of traditional experiments when revealing drug-protein interactions (DTIs), the screening of targets not only takes a lot of time and money but also has high false-positive and false-negative rates. Therefore, it is imperative to develop effective automatic computational methods to accurately predict DTIs in the postgenome era. In this article, we propose a new computational method for predicting DTIs from drug molecular structure and protein sequence by using the stacked autoencoder of deep learning, which can adequately extract the raw data information. The proposed method has the advantage that it can automatically mine the hidden information from protein sequences and generate highly representative features through iterations of multiple layers. The feature descriptors are then constructed by combining the molecular substructure fingerprint information, and fed into the rotation forest for accurate prediction. The experimental results of fivefold cross-validation indicate that the proposed method achieves superior performance on gold standard data sets (enzymes, ion channels, GPCRs [G-protein-coupled receptors], and nuclear receptors) with accuracy of 0.9414, 0.9116, 0.8669, and 0.8056, respectively. We further comprehensively explore the performance of the proposed method by comparing it with other feature extraction algorithms, state-of-the-art classifiers, and other excellent methods on the same data set. The excellent comparison results demonstrate that the proposed method is highly competitive when predicting drug-target interactions.

Unsupervised deep learning with variational autoencoders applied to breast tumor genome-wide DNA methylation data with biologic feature extraction

“…Our initial effort to use deep learning with breast tumor DNA methylation data to recapitulate intrinsic breast tumor molecular subtypes established the feasibility and promise of unsupervised deep learning approaches for molecular data sets [11]. Other work has begun to use unsupervised deep learning to investigate shared transcriptomic signatures in a pan-cancer setting [10], and to identify a potential response to drug treatments and drug-drug interactions [5,6,7,8]. There is potential, then, to use latent representations of tumors in this way to predict a patient's response to therapies, or to stratify patients by latent risk profile.…”

“…This in turn provides an opportunity to study genome-wide interactions in ways that may be computationally intractable using traditional statistical modeling approaches. Indeed, to date, deep learning have been used to generate photo-realistic cell images [3], learn functional representations of neural in-situ hybridization images [4], to predict novel drug targets [5,6,7,8], and to model the hierarchical structure and function of the cell [9]. Recently, VAE methods have also been shown to learn biologically-relevant latent representations of tumors using genome-scale gene expression [10] and DNA methylation data11.…”

Computational prediction of drug–target interactions using chemogenomic approaches: an empirical survey

“…A number of efforts regarding DTI prediction have made use of deep learning to improve prediction performance. Deep learning techniques that have been used in DTI prediction include restricted Boltzmann machines [114], deep neural networks [115,116,117], stacked auto-encoders [118,119] and deep belief networks [120]. As of yet, none of the deep learning methods developed for DTI prediction have attempted to simultaneously use multiple heterogeneous sources of drug and target information.…”

RFDTI: Using Rotation Forest with Feature Weighted for Drug-Target Interaction Prediction from Drug Molecular Structure and Protein Sequence

“…In the experiment, we used Position-Specific Scoring Matrix (PSSM) to convert protein sequence numerically [30]. PSSM is widely used in protein binding site prediction, protein secondary structure prediction, and prediction of disordered regions .…”