Drug-target Prediction Research Articles

ArkDTA: attention regularization guided by non-covalent interactions for explainable drug-target binding affinity prediction.

Protein-ligand binding affinity prediction is a central task in drug design and development. Cross-modal attention mechanism has recently become a core component of many deep learning models due to its potential to improve model explainability. Non-covalent interactions (NCIs), one of the most critical domain knowledge in binding affinity prediction task, should be incorporated into protein-ligand attention mechanism for more explainable deep drug-target interaction models. We propose ArkDTA, a novel deep neural architecture for explainable binding affinity prediction guided by NCIs. Experimental results show that ArkDTA achieves predictive performance comparable to current state-of-the-art models while significantly improving model explainability. Qualitative investigation into our novel attention mechanism reveals that ArkDTA can identify potential regions for NCIs between candidate drug compounds and target proteins, as well as guiding internal operations of the model in a more interpretable and domain-aware manner. ArkDTA is available at https://github.com/dmis-lab/ArkDTA. kangj@korea.ac.kr.

DrugOOD: Out-of-Distribution Dataset Curator and Benchmark for AI-Aided Drug Discovery – a Focus on Affinity Prediction Problems with Noise Annotations

AI-aided drug discovery (AIDD) is gaining popularity due to its potential to make the search for new pharmaceuticals faster, less expensive, and more effective. Despite its extensive use in numerous fields (e.g., ADMET prediction, virtual screening), little research has been conducted on the out-of-distribution (OOD) learning problem with noise. We present DrugOOD, a systematic OOD dataset curator and benchmark for AIDD. Particularly, we focus on the drug-target binding affinity prediction problem, which involves both macromolecule (protein target) and small-molecule (drug compound). DrugOOD offers an automated dataset curator with user-friendly customization scripts, rich domain annotations aligned with biochemistry knowledge, realistic noise level annotations, and rigorous benchmarking of SOTA OOD algorithms, as opposed to only providing fixed datasets. Since the molecular data is often modeled as irregular graphs using graph neural network (GNN) backbones, DrugOOD also serves as a valuable testbed for graph OOD learning problems. Extensive empirical studies have revealed a significant performance gap between in-distribution and out-of-distribution experiments, emphasizing the need for the development of more effective schemes that permit OOD generalization under noise for AIDD.

Differential Expression of MicroRNAs and Predicted Drug Target in Amyotrophic Lateral Sclerosis.

ALS (Amyotrophic Lateral Sclerosis) is a rare type of neurodegenerative disease. It shows progressive degradation of motor neurons in the brain and spinal cord. At present, there is no treatment available that can completely cure ALS. The available treatments can only increase a patient's life span by a few months. Recently, microRNAs (miRNAs), a sub-class of small non-coding RNAs have been shown to play an essential role in the diagnosis, prognosis, and therapy of ALS. Our study focuses on analyzing differential miRNA profiles and predicting drug targets in ALS using bioinformatics and computational approach. The study identifies eight highly differentially expressed miRNAs in ALS patients, four of which are novel. We identified 42 hub genes for these eight highly expressed miRNAs with Amyloid Precursor Protein (APP) as a candidate gene among them for highly expressed down-regulated miRNA, hsa-miR-455-3p using protein-protein interaction network and Cytoscape analysis. A novel association has been found between hsa-miR-455-3p/APP/serotonergic pathway using KEGG pathway analysis. Also, molecular docking studies have revealed curcumin as a potential drug target that may be used for the treatment of ALS. Thus, the present study has identified four novel miRNA biomarkers: hsa-miR-3613-5p, hsa-miR-24, hsa-miR-3064-5p, and hsa-miR-4455. There is a formation of a novel axis, hsa-miR-455-3p/APP/serotonergic pathway, and curcumin is predicted as a potential drug target for ALS.

NHGNN-DTA: a node-adaptive hybrid graph neural network for interpretable drug-target binding affinity prediction.

Large-scale prediction of drug-target affinity (DTA) plays an important role in drug discovery. In recent years, machine learning algorithms have made great progress in DTA prediction by utilizing sequence or structural information of both drugs and proteins. However, sequence-based algorithms ignore the structural information of molecules and proteins, while graph-based algorithms are insufficient in feature extraction and information interaction. In this article, we propose NHGNN-DTA, a node-adaptive hybrid neural network for interpretable DTA prediction. It can adaptively acquire feature representations of drugs and proteins and allow information to interact at the graph level, effectively combining the advantages of both sequence-based and graph-based approaches. Experimental results have shown that NHGNN-DTA achieved new state-of-the-art performance. It achieved the mean squared error (MSE) of 0.196 on the Davis dataset (below 0.2 for the first time) and 0.124 on the KIBA dataset (3% improvement). Meanwhile, in the case of cold start scenario, NHGNN-DTA proved to be more robust and more effective with unseen inputs than baseline methods. Furthermore, the multi-head self-attention mechanism endows the model with interpretability, providing new exploratory insights for drug discovery. The case study on Omicron variants of SARS-CoV-2 illustrates the efficient utilization of drug repurposing in COVID-19. The source code and data are available at https://github.com/hehh77/NHGNN-DTA.

Prognostic immunogenic characteristics of iron pendant disease modifiers in colon cancer.

We explored the prognostic and immunogenic characteristics of iron pendant disease regulators in colon cancer to provide a scientific basis for the prediction of tumor prognosis-related markers and potential immunotherapeutic drug targets. RNA sequencing and matched complete clinical information of colon cancer (COAD) were retrieved from the UCSC Xena database, and genomic and transcriptomic data of colon cancer from the TCGA database were downloaded. Then univariate and multifactorial Cox regression were used to process these data. The prognostic factors were analyzed by single-factor and multi-factor Cox regression, followed by Kaplan-Meier survival curves with the aid of R software "survival" package. Then we use FireBrowse online analysis tool to analyze the expression variation of all cancer genes, and draw a histogram according to the influencing factors to predict the 1, 3, and 5 year survival rates of patients. The results show that age, tumor stage and iron death score were significantly correlated with prognosis (p<0.05). Further multivariate cox regression analysis confirmed that age, tumor stage and iron death score were still significantly correlated with prognosis (p<0.05); The calibration curve results show that the deviation between the predicted values of 1 year, 3 years and 5 years and the diagonal of the figure is very small; the ROC curve results show that the AUC values of the 1-year and 5-year ROC curves of the bar graph are high; the DCA curve results show that the net yield of the bar graph is the largest; The scores of T cells and B cells in the high iron death score group were significantly lower than those in the low iron death score group, and the activities of immune related pathways were significantly reduced. There was a significant difference in the iron death score between the iron death molecular subtype and the gene cluster subtype. The model showed a superior response to immunotherapy in the high-risk group, revealing a potential relationship between iron death and tumor immunotherapy, which will provide new ideas for the treatment and prognostic assessment of colon cancer patients.

Design and development of salen-type Schiff bases as potential antivirus agents: Experimental and theoretical approach

Five new salen-type Schiff base compounds (LH2-L4H2) have been designed and synthesized and their interaction with SARS-CoV-2, HIV virus and DNA polymerase IV have been studied by in&nbsp;silico approaches. The newly synthesized salen-type Schiff base ligands have been derived from the condensation of substituted salicylaldehydes and ethylenediamine in MeOH. 1H and 13C&nbsp;NMR, FT-IR and UV-Vis spectral techniques have been applied in order to confirm the structural elucidation of the desired products. The crystal structure of the compound LH2 is determined by the single crystal X-ray diffraction. Density functional theory (DFT) calculations have been employed to evaluate the optimized electronic structure, HOMO-LUMO, energy gap, and global parameters. Extensive classical molecular dynamics simulations have been performed to investigate the consequences of docking of the synthesized ligand, LH2 on a protein moiety (PDB id: 7O46). Molecular docking studies have been performed on compounds (LH2-L4H2) to predict the binding mode and interactions between the ligands and the main protease of the SARS-CoV-2 (PDB ID: 7O46) for COVID-19 and HIV virus (PDB ID: 1UUI). The molecular docking results show that compounds (LH2-L4H2) with SARS-CoV-2 and HIV virus exhibit good binding affinity at binding site of receptor protein. As potential drug candidates, Swiss-ADME and target predictions (pharmacokinetics and drug-likeness prediction) analyses have also been studied and the results are compared with Chloroquine (CQ) and Hydroxychloroquine (HCQ) as anti-SARS-CoV-2 drugs. Salen-type ligands have also been docked into the DNA polymerase IV (PDB ID: 5YUX) for surface binding and intercalation. Overall, the present study offers the therapeutic potential for a series of compounds (LH2-L4H2) for biomedical applications with reference to coronavirus (SARS-CoV-2) and HIV virus.

Predicting drug adverse effects using a new Gastro-Intestinal Pacemaker Activity Drug Database (GIPADD)

Electrical data could be a new source of big-data for training artificial intelligence (AI) for drug discovery. A Gastro-Intestinal Pacemaker Activity Drug Database (GIPADD) was built using a standardized methodology to test drug effects on electrical gastrointestinal (GI) pacemaker activity. The current report used data obtained from 89 drugs with 4867 datasets to evaluate the potential use of the GIPADD for predicting drug adverse effects (AEs) using a machine-learning (ML) approach and to explore correlations between AEs and GI pacemaker activity. Twenty-four “electrical” features (EFs) were extracted using an automated analytical pipeline from the electrical signals recorded before and after acute drug treatment at three concentrations (or more) on four-types of GI tissues (stomach, duodenum, ileum and colon). Extracted features were normalized and merged with an online side-effect resource (SIDER) database. Sixty-six common AEs were selected. Different algorithms of classification ML models, including Naïve Bayes, discriminant analysis, classification tree, k-nearest neighbors, support vector machine and an ensemble model were tested. Separated tissue models were also tested. Averaging experimental repeats and dose adjustment were performed to refine the prediction results. Random datasets were created for model validation. After model validation, nine AEs classification ML model were constructed with accuracy ranging from 67 to 80%. EF can be further grouped into ‘excitatory’ and ‘inhibitory’ types of AEs. This is the first time drugs are being clustered based on EF. Drugs acting on similar receptors share similar EF profile, indicating potential use of the database to predict drug targets too. GIPADD is a growing database, where prediction accuracy is expected to improve. The current approach provides novel insights on how EF may be used as new source of big-data in health and disease.

Improving drug-target affinity prediction via feature fusion and knowledge distillation.

Rapid and accurate prediction of drug-target affinity can accelerate and improve the drug discovery process. Recent studies show that deep learning models may have the potential to provide fast and accurate drug-target affinity prediction. However, the existing deep learning models still have their own disadvantages that make it difficult to complete the task satisfactorily. Complex-based models rely heavily on the time-consuming docking process, and complex-free models lacks interpretability. In this study, we introduced a novel knowledge-distillation insights drug-target affinity prediction model with feature fusion inputs to make fast, accurate and explainable predictions. We benchmarked the model on public affinity prediction and virtual screening dataset. The results show that it outperformed previous state-of-the-art models and achieved comparable performance to previous complex-based models. Finally, we study the interpretability of this model through visualization and find it can provide meaningful explanations for pairwise interaction. We believe this model can further improve the drug-target affinity prediction for its higher accuracy and reliable interpretability.

DrugormerDTI: Drug Graphormer for drug–target interaction prediction

Drug-target interactions (DTI) prediction is a crucial task in drug discovery. Existing computational methods accelerate the drug discovery in this respect. However, most of them suffer from low feature representation ability, significantly affecting the predictive performance. To address the problem, we propose a novel neural network architecture named DrugormerDTI, which uses Graph Transformer to learn both sequential and topological information through the input molecule graph and Resudual2vec to learn the underlying relation between residues from proteins. By conducting ablation experiments, we verify the importance of each part of the DrugormerDTI. We also demonstrate the good feature extraction and expression capabilities of our model via comparing the mapping results of the attention layer and molecular docking results. Experimental results show that our proposed model performs better than baseline methods on four benchmarks. We demonstrate that the introduction of Graph Transformer and the design of residue are appropriate for drug–target prediction.

Abstract 3149: Prediction of metabolic drug targets for cancer patients with poor prognosis by using genome-scale metabolic models and survival analyses

Abstract Cancer metabolism has been extensively explored to identify biomarkers and novel drug targets for effective cancer treatment. In particular, it is expected that metabolism-derived data somehow reflects unique features associated with a cancer patient’s prognosis. Here, we introduce a computational framework that predicts metabolic drug targets by considering metabolites associated with poor prognosis. For inputs, the computational framework requires genome-scale metabolic models (GEMs) and survival data that represent each of cancer patients with diverse prognoses. GEM is a computational model that allows predicting fluxes of entire metabolic reactions in a patient-specific manner. By using this computational framework, a risk score, which reflects the prognosis of cancer patients, was first established on the basis of metabolites significantly associated with survival. Finally, metabolic reaction targets were predicted that could reduce the risk score (i.e., improving the prognosis) upon their knockdown in silico. In this study, this framework was applied to bladder cancer patients as a proof-of-concept demonstration, and predicted 13 metabolic reaction targets. These metabolic reaction targets were expected to be also effective to bladder cancer patients with poor prognosis. The computational framework developed in this study is distinct from previous drug targeting methods because it also considers unique metabolic features associated with cancer patients with poor prognosis in order to identify drug targets that are effective to both low-risk and high-risk groups of cancer patients. Citation Format: Sang Mi Lee, GaRyoung Lee, Hyun Uk Kim. Prediction of metabolic drug targets for cancer patients with poor prognosis by using genome-scale metabolic models and survival analyses [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3149.

BiComp-DTA: Drug-target binding affinity prediction through complementary biological-related and compression-based featurization approach

Drug-target binding affinity prediction plays a key role in the early stage of drug discovery. Numerous experimental and data-driven approaches have been developed for predicting drug-target binding affinity. However, experimental methods highly rely on the limited structural-related information from drug-target pairs, domain knowledge, and time-consuming assays. On the other hand, learning-based methods have shown an acceptable prediction performance. However, most of them utilize several simple and complex types of proteins and drug compounds data, ranging from the protein sequences to the topology of a graph representation of drug compounds, employing multiple deep neural networks for encoding and feature extraction, and so, leads to the computational overheads. In this study, we propose a unified measure for protein sequence encoding, named BiComp, which provides compression-based and evolutionary-related features from the protein sequences. Specifically, we employ Normalized Compression Distance and Smith-Waterman measures for capturing complementary information from the algorithmic information theory and biological domains, respectively. We utilize the proposed measure to encode the input proteins feeding a new deep neural network-based method for drug-target binding affinity prediction, named BiComp-DTA. BiComp-DTA is evaluated utilizing four benchmark datasets for drug-target binding affinity prediction. Compared to the state-of-the-art methods, which employ complex models for protein encoding and feature extraction, BiComp-DTA provides superior efficiency in terms of accuracy, runtime, and the number of trainable parameters. The latter achievement facilitates execution of BiComp-DTA on a normal desktop computer in a fast fashion. As a comparative study, we evaluate BiComp's efficiency against its components for drug-target binding affinity prediction. The results have shown superior accuracy of BiComp due to the orthogonality and complementary nature of Smith-Waterman and Normalized Compression Distance measures for protein sequences. Such a protein sequence encoding provides efficient representation with no need for multiple sources of information, deep domain knowledge, and complex neural networks.

A graph-based multilevel property extraction network for predicting protein drug target binding affinity

Background: As the basic material of life, protein regulates physiological activities such as material in and out of cells, signal transduction, metabolism and so on. However, studies have shown that proteins cannot perform these functions alone in cells, and need to be combined with ligands to perform functions. background: As the basic material of life, protein regulates physiological activities such as material in and out of cells, signal transduction, metabolism and so on. However, studies have shown that proteins cannot perform these functions alone in cells, and need to be combined with ligands to perform functions. Objective: The discovery of drugs is based on this mechanism to be developed. But at present, it is difficult to discover new drugs through biological experiments, which lead to high cost, deteriorate the environment and make the human body more resistant to drugs. objective: The discovery of drugs is based on this mechanism to be developed. But at present, it is difficult to discover new drugs through biological experiments, which lead to cost high, deteriorate the environment and make human body more resistant to drugs. The rapid development of computer can assist researchers to screen out potential drugs that can bind to proteins in advance. In the past few decades, most of drug-target affinity prediction methods have disadvantages in high requirements on data set and difficulties to predict the binding strength of drug-target. Method: The rapid development of computers can assist researchers in screening potential drugs that can bind to proteins in advance. In the past few decades, most of the protein drug-target affinity prediction methods have had disadvantages of high requirements on data set and difficulties to predict the binding strength of drug-target. method: This paper proposes a multi-level feature extraction model based on graph convolution to predict drug-target affinity. The model uses an integrated neural network of CNN and GNN to learn the characteristics of the input data of drug-targets. Results: This paper proposes a multilevel feature extraction model based on graph convolution to predict protein drug-target affinity. The model uses an integrated neural network of Text CNN and GNN to learn the characteristics of the input data of drug-targets. result: Experimental results on the benchmark datasets of Davis and Kiba showed that the proposed graph-based convolution network achieves good performance on drug-target affinity prediction. Conclusion: Experimental results on the benchmark datasets of Davis and Kiba showed that the proposed graph-based convolution network achieves good performance on drug-target affinity prediction. conclusion: The results showed that the multi-level feature extraction network model based on graph convolution (GCN-GAT-GCN) is more effective in learning molecular feature information. other: None

Rheumatoid arthritis treatment with zoledronic acid, a potential inhibitor of GWAS-derived pharmacogenetics STAT3 and IL2 targets

Rheumatoid arthritis (RA) is an inflammatory condition that primarily affects the joints and progress to affect other vital organs. Variety of drugs are being recommended to control the disease progression that benefits patients to perform day-to-day activities. Few of these RA drugs have noticeable side effects; therefore, it's crucial to choose the appropriate drug for treating RA with an understanding of the disease's pathophysiology. Herein, we investigated the RA genes from GWAS data to construct protein–protein interaction (PPI) network and to define appropriate drug targets for RA. The predicted drug targets were screened with the known RA drugs based on molecular docking. Further, the molecular dynamics simulations were performed to comprehend the conformational changes and stability of the targets upon binding of the selected top ranked RA drug. As a result, our constructed protein network from GWAS data revealed, STAT3 and IL2 could be potential pharmacogenetics targets that interlink most of the RA genes encoding proteins. These interlinked proteins of both the targets showed involvement in cell signaling, immune response, and TNF signaling pathway. Among the 192 RA drugs investigated, zoledronic acid had the lowest binding energy that inhibit both STAT3 (-6.307 kcal/mol) and IL2 (-6.231 kcal/mol). Additionally, STAT3 and IL2 trajectories on zoledronic acid binding exhibit notable differences in MD simulations as compared to a drug-free environment. Also, the in vitro assessment with the zoledronic acid confirms the outcome of our computational study. Overall, our study identify zoledronic acid could be potential inhibitor against these targets, that will benefits patients with RA. Comparative efficiency assessments between the RA drugs through clinical trials are needed to validate our findings in the treatment of RA.

Drug–target interaction prediction based on protein features, using wrapper feature selection

Drug–target interaction prediction is a vital stage in drug development, involving lots of methods. Experimental methods that identify these relationships on the basis of clinical remedies are time-taking, costly, laborious, and complex introducing a lot of challenges. One group of new methods is called computational methods. The development of new computational methods which are more accurate can be preferable to experimental methods, in terms of total cost and time. In this paper, a new computational model to predict drug–target interaction (DTI), consisting of three phases, including feature extraction, feature selection, and classification is proposed. In feature extraction phase, different features such as EAAC, PSSM and etc. would be extracted from sequence of proteins and fingerprint features from drugs. These extracted features would then be combined. In the next step, one of the wrapper feature selection methods named IWSSR, due to the large amount of extracted data, is applied. The selected features are then given to rotation forest classification, to have a more efficient prediction. Actually, the innovation of our work is that we extract different features; and then select features by the use of IWSSR. The accuracy of the rotation forest classifier based on tenfold on the golden standard datasets (enzyme, ion channels, G-protein-coupled receptors, nuclear receptors) is as follows: 98.12, 98.07, 96.82, and 95.64. The results of experiments indicate that the proposed model has an acceptable rate in DTI prediction and is compatible with the proposed methods in other papers.

EGCG identified as an autophagy inducer for rosacea therapy.

Background: Rosacea is a common facial skin inflammatory disease featured by hyperactivation of mTORC1 signaling in the epidermis. Due to unclear pathogenesis, the effective treatment options for rosacea remain limited. Methods: Weighted gene co-expression network analysis (WGCNA) analyzed the relationship between epidermis autophagy and mTOR pathways in rosacea, and further demonstrated it through immunofluorescence and qPCR analysis. A potential therapeutic agent for rosacea was predicted based on the key genes of the WGCNA module. In vivo and in vitro experiments were conducted to verify its therapeutic role. Drug-target prediction (TargetNet, Swiss, and Tcmsp) and molecular docking offered potential pharmacological targets. Results: WGCNA showed that epidermis autophagy was related to the activation of mTOR pathways in rosacea. Next, autophagy was downregulated in the epidermis of rosacea, which was regulated by mTOR. In addition, the in vivo experiment demonstrated that autophagy induction could be an effective treatment strategy for rosacea. Subsequently, based on the key genes of the WGCNA module, epigallocatechin-3-gallate (EGCG) was predicted as a potential therapeutic agent for rosacea. Furthermore, the therapeutic role of EGCG on rosacea was confirmed in vivo and in vitro. Finally, drug-target prediction and molecular docking revealed that AKT1/MAPK1/MMP9 could be the pharmacological targets of EGCG in rosacea. Conclusion: Collectively, our findings revealed the vital role of autophagy in rosacea and identified that EGCG, as a therapeutic agent for rosacea, attenuated rosacea-like inflammation via inducing autophagy in keratinocytes.

Prediction of drug protein interactions based on variable scale characteristic pyramid convolution network.

In the process of drug screening, it is significant to improve the accuracy of drug-target binding affinity prediction. A multilayer convolutional neural network is one of the most popular existing methods for predicting affinity based on deep learning. It uses multiple convolution layers to extract features from the simplified molecular input system (SMILES) strings of the compounds and amino acid sequences of proteins and then performs affinity prediction analysis. However, the semantic information contained in low-level features can gradually be lost due to the increasing network depth, which affects the prediction performance. We propose a novel method called the Pyramid Network Convolution Drug-Target Binding Affinity (PCNN-DTA) method for drug-target binding affinity prediction. The PCNN-DTA method, which is based on a feature pyramid network (FPN), fuses the features extracted from each layer of a multilayer convolution network to retain more low-level feature information, thus improving the prediction accuracy. PCNN-DTA is compared with other typical algorithms on three benchmark datasets, namely, the KIBA, Davis, and Binding DB datasets. Experimental results show that the PCNN-DTA method is superior to existing regression prediction methods using convolutional neural networks, which further demonstrates its effectiveness.

Explaining Black Box Drug Target Prediction Through Model Agnostic Counterfactual Samples.

Many high-performance DTA deep learning models have been proposed, but they are mostly black-box and thus lack human interpretability. Explainable AI (XAI) can make DTA models more trustworthy, and allows to distill biological knowledge from the models. Counterfactual explanation is one popular approach to explaining the behaviour of a deep neural network, which works by systematically answering the question "How would the model output change if the inputs were changed in this way?". We propose a multi-agent reinforcement learning framework, Multi-Agent Counterfactual Drug-target binding Affinity (MACDA), to generate counterfactual explanations for the drug-protein complex. Our proposed framework provides human-interpretable counterfactual instances while optimizing both the input drug and target for counterfactual generation at the same time. We benchmark the proposed MACDA framework using the Davis and PDBBind dataset and find that our framework produces more parsimonious explanations with no loss in explanation validity, as measured by encoding similarity. We then present a case study involving ABL1 and Nilotinib to demonstrate how MACDA can explain the behaviour of a DTA model in the underlying substructure interaction between inputs in its prediction, revealing mechanisms that align with prior domain knowledge.

AttentionDTA: Drug-Target Binding Affinity Prediction by Sequence-Based Deep Learning With Attention Mechanism.

The identification of drug-target relations (DTRs) is substantial in drug development. A large number of methods treat DTRs as drug-target interactions (DTIs), a binary classification problem. The main drawback of these methods are the lack of reliable negative samples and the absence of many important aspects of DTR, including their dose dependence and quantitative affinities. With increasing number of publications of drug-protein binding affinity data recently, DTRs prediction can be viewed as a regression problem of drug-target affinities (DTAs) which reflects how tightly the drug binds to the target and can present more detailed and specific information than DTIs. The growth of affinity data enables the use of deep learning architectures, which have been shown to be among the state-of-the-art methods in binding affinity prediction. Although relatively effective, due to the black-box nature of deep learning, these models are less biologically interpretable. In this study, we proposed a deep learning-based model, named AttentionDTA, which uses attention mechanism to predict DTAs. Different from the models using 3D structures of drug-target complexes or graph representation of drugs and proteins, the novelty of our work is to use attention mechanism to focus on key subsequences which are important in drug and protein sequences when predicting its affinity. We use two separate one-dimensional Convolution Neural Networks (1D-CNNs) to extract the semantic information of drug's SMILES string and protein's amino acid sequence. Furthermore, a two-side multi-head attention mechanism is developed and embedded to our model to explore the relationship between drug features and protein features. We evaluate our model on three established DTA benchmark datasets, Davis, Metz, and KIBA. AttentionDTA outperforms the state-of-the-art deep learning methods under different evaluation metrics. The results show that the attention-based model can effectively extract protein features related to drug information and drug features related to protein information to better predict drug target affinities. It is worth mentioning that we test our model on IC50 dataset, which provides the binding sites between drugs and proteins, to evaluate the ability of our model to locate binding sites. Finally, we visualize the attention weight to demonstrate the biological significance of the model. The source code of AttentionDTA can be downloaded from https://github.com/zhaoqichang/AttentionDTA_TCBB.

Computer Assisted Drug Design of AITC for Treatment of Cancer: A Combined Molecular Target Identification Study and Drug-Target Prediction

Drugs can influence the whole biological system by targeting interaction reactions. However, current databases have minimal drug target information and only a few are validated. This emerges a need to develop new drugs that predict true interaction pairs with high accuracy. Currently, network pharmacology has been used in identifying potential drug targets for allylisothiocyanate (AITC) and in predicting the spread of its activity toward bladder cancer whereas the mechanisms of isothiocyanates remain largely undefined. In this study, we present a computational method to predict targets for AITC. Results showed 13 differentially expressed hubs confirmed by Cytoscape (3.7.2). We further observed that the obtained targets were found to involve a type of Ontology of classification- molecular function, cellular component, biological process and protein class. Pathway enrichment was done by DAVID. This study identifies key genes and pathways involved in bladder cancer which helps us to understand the progression and molecular mechanism underlying bladder cancer. The obtained key genes may serve as active therapeutic targets for bladder cancer treatment. AITC may directly regulate several biological processes and their genes in bladder cancer.

A machine learning strategy with clustering under sampling of majority instances for predicting drug target interactions.

Drug Target Interactions (DTIs) are crucial in drug discovery as it reduces the range of candidate searches, speeding up the drug screening process. Considering in vitro and in vivo experimentations are time and cost-expensive, there has been a surge in computational techniques, especially ML methods for DTIs prediction. Therefore, this study aims to present a methodology that uses molecular structures and amino acid sequences for generating PSSM and PubChem fingerprints for drugs and targets respectively. The proposed work uses a novel technique NearestCUS for handling the class imbalance problem of the benchmark datasets. We use Isomap Embedding to extract features from PSSMs. Feature selection is performed using ANOVA. CatBoost is used for predicting the interaction between drugs and targets for the first time. To quantify the efficacy of NearestCUS, we compared it with other sampling techniques. We found that the proposed methodology performed better than state-of-the-art approaches.