Artificial intelligence in drug discovery and development

Drug Discovery Today, Journal Year: 2020, Volume and Issue: 26(1), P. 80 - 93

Published: Oct. 21, 2020

Language: Английский

---

Artificial intelligence to deep learning: machine intelligence approach for drug discovery

Molecular Diversity, Journal Year: 2021, Volume and Issue: 25(3), P. 1315 - 1360

Published: April 12, 2021

Drug designing and development is an important area of research for pharmaceutical companies chemical scientists. However, low efficacy, off-target delivery, time consumption, high cost impose a hurdle challenges that impact drug design discovery. Further, complex big data from genomics, proteomics, microarray data, clinical trials also obstacle in the discovery pipeline. Artificial intelligence machine learning technology play crucial role development. In other words, artificial neural networks deep algorithms have modernized area. Machine been implemented several processes such as peptide synthesis, structure-based virtual screening, ligand-based toxicity prediction, monitoring release, pharmacophore modeling, quantitative structure-activity relationship, repositioning, polypharmacology, physiochemical activity. Evidence past strengthens implementation this field. Moreover, novel mining, curation, management techniques provided critical support to recently developed modeling algorithms. summary, advancements provide excellent opportunity rational process, which will eventually mankind. The primary concern associated with consumption production cost. inefficiency, inaccurate target inappropriate dosage are hurdles inhibit process delivery With technology, computer-aided integrating can eliminate traditional referred superset comprising learning, whereas comprises supervised unsupervised reinforcement learning. subset has extensively network, vector machines, classification regression, generative adversarial networks, symbolic meta-learning examples applied process. different areas synthesis molecule design, screening molecular docking, relationship protein misfolding protein-protein interactions, pathway identification polypharmacology. principles active inactive, pre-clinical development, secondary biomarker manufacturing, bioactivity properties, prediction toxicity, mode action.

Language: Английский

---

Predicting commercially available antiviral drugs that may act on the novel coronavirus (SARS-CoV-2) through a drug-target interaction deep learning model

Computational and Structural Biotechnology Journal, Journal Year: 2020, Volume and Issue: 18, P. 784 - 790

Published: Jan. 1, 2020

The infection of a novel coronavirus found in Wuhan China (SARS-CoV-2) is rapidly spreading, and the incidence rate increasing worldwide. Due to lack effective treatment options for SARS-CoV-2, various strategies are being tested China, including drug repurposing. In this study, we used our pre-trained deep learning-based drug-target interaction model called Molecule Transformer-Drug Target Interaction (MT-DTI) identify commercially available drugs that could act on viral proteins SARS-CoV-2. result showed atazanavir, an antiretroviral medication treat prevent human immunodeficiency virus (HIV), best chemical compound, showing inhibitory potency with Kd 94.94 nM against SARS-CoV-2 3C-like proteinase, followed by remdesivir (113.13 nM), efavirenz (199.17 ritonavir (204.05 dolutegravir (336.91 nM). Interestingly, lopinavir, ritonavir, darunavir all designed target proteinases. However, prediction, they may also bind replication complex components < 1000 nM. addition, several antiviral agents, such as Kaletra (lopinavir/ritonavir), be Overall, suggest list identified MT-DTI should considered, when establishing

Language: Английский

---

Drug discovery with explainable artificial intelligence

Nature Machine Intelligence, Journal Year: 2020, Volume and Issue: 2(10), P. 573 - 584

Published: Oct. 13, 2020

Language: Английский

---

GraphDTA: predicting drug–target binding affinity with graph neural networks

Bioinformatics, Journal Year: 2020, Volume and Issue: 37(8), P. 1140 - 1147

Published: Oct. 15, 2020

Abstract Summary The development of new drugs is costly, time consuming and often accompanied with safety issues. Drug repurposing can avoid the expensive lengthy process drug by finding uses for already approved drugs. In order to repurpose effectively, it useful know which proteins are targeted Computational models that estimate interaction strength drug–target pairs have potential expedite repurposing. Several been proposed this task. However, these represent as strings, not a natural way molecules. We propose model called GraphDTA represents graphs graph neural networks predict affinity. show only affinity better than non-deep learning models, but also outperform competing deep methods. Our results confirm appropriate binding prediction, representing lead further improvements. Availability implementation implemented in Python. Related data, pre-trained source code publicly available at https://github.com/thinng/GraphDTA. All scripts data needed reproduce post hoc statistical analysis from https://doi.org/10.5281/zenodo.3603523. Supplementary information Bioinformatics online.

Language: Английский

---

DeepConv-DTI: Prediction of drug-target interactions via deep learning with convolution on protein sequences

PLoS Computational Biology, Journal Year: 2019, Volume and Issue: 15(6), P. e1007129 - e1007129

Published: June 14, 2019

Identification of drug-target interactions (DTIs) plays a key role in drug discovery. The high cost and labor-intensive nature vitro vivo experiments have highlighted the importance silico-based DTI prediction approaches. In several computational models, conventional protein descriptors been shown to not be sufficiently informative predict accurate DTIs. Thus, this study, we propose deep learning based model capturing local residue patterns proteins participating When employ convolutional neural network (CNN) on raw sequences, perform convolution various lengths amino acids subsequences capture generalized classes. We train our with large-scale information demonstrate performance proposed using an independent dataset that is seen during training phase. As result, performs better than previous descriptor-based models. Also, recently developed models for massive By examining pooled results, confirmed can detect binding sites conclusion, detecting target successfully enriches features sequence, yielding results Our code available at https://github.com/GIST-CSBL/DeepConv-DTI.

Language: Английский

---

TCM network pharmacology: A new trend towards combining computational, experimental and clinical approaches

Chinese Journal of Natural Medicines, Journal Year: 2021, Volume and Issue: 19(1), P. 1 - 11

Published: Jan. 1, 2021

Language: Английский

---

Predicting Drug–Target Interaction Using a Novel Graph Neural Network with 3D Structure-Embedded Graph Representation

Journal of Chemical Information and Modeling, Journal Year: 2019, Volume and Issue: 59(9), P. 3981 - 3988

Published: Aug. 23, 2019

We propose a novel deep learning approach for predicting drug-target interaction using graph neural network. introduce distance-aware attention algorithm to differentiate various types of intermolecular interactions. Furthermore, we extract the feature interactions directly from 3D structural information on protein-ligand binding pose. Thus, model can learn key features accurate predictions rather than just memorize certain patterns ligand molecules. As result, our shows better performance docking and other methods both virtual screening (AUROC 0.968 DUD-E test set) pose prediction 0.935 PDBbind set). In addition, it reproduce natural population distribution active molecules inactive

Language: Английский

---

Machine learning approaches and databases for prediction of drug–target interaction: a survey paper

Briefings in Bioinformatics, Journal Year: 2019, Volume and Issue: 22(1), P. 247 - 269

Published: Nov. 8, 2019

Abstract The task of predicting the interactions between drugs and targets plays a key role in process drug discovery. There is need to develop novel efficient prediction approaches order avoid costly laborious yet not-always-deterministic experiments determine drug–target (DTIs) by alone. These should be capable identifying potential DTIs timely manner. In this article, we describe data required for DTI followed comprehensive catalog consisting machine learning methods databases, which have been proposed utilized predict DTIs. advantages disadvantages each set are also briefly discussed. Lastly, challenges one may face using highlighted conclude shedding some lights on important future research directions.

Language: Английский

---

DeepPurpose: a deep learning library for drug–target interaction prediction

Bioinformatics, Journal Year: 2020, Volume and Issue: 36(22-23), P. 5545 - 5547

Published: Nov. 20, 2020

Abstract Summary Accurate prediction of drug–target interactions (DTI) is crucial for drug discovery. Recently, deep learning (DL) models show promising performance DTI prediction. However, these can be difficult to use both computer scientists entering the biomedical field and bioinformaticians with limited DL experience. We present DeepPurpose, a comprehensive easy-to-use library DeepPurpose supports training customized by implementing 15 compound protein encoders over 50 neural architectures, along providing many other useful features. demonstrate state-of-the-art on several benchmark datasets. Availability implementation https://github.com/kexinhuang12345/DeepPurpose. Supplementary information data are available at Bioinformatics online.

Language: Английский

---