SAG-DTA: Prediction of Drug–Target Affinity Using Self-Attention Graph Network

Zhang, Shuguang; Jiang, Mingjian; Wang, Shuang; Wei, Zhiqiang; Li, Zhen

doi:10.3390/ijms22168993

Cited by 35 publications

(12 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This kind of incremental parameter design can enhance the information transfer between atoms. SAGDTA [ 44 ]: It exploited the self-attention mechanism on drug molecular graphs to obtain efficient representations of drugs. In this study, features of each atom node in the molecular graph and the SAG used the hierarchical pooing architecture with 3 blocks which has been demonstrated to absorb global information better.…”

Section: Resultsmentioning

confidence: 99%

GeneralizedDTA: combining pre-training and multi-task learning to predict drug-target binding affinity for unknown drug discovery

2022

View full text Add to dashboard Cite

Background Accurately predicting drug-target binding affinity (DTA) in silico plays an important role in drug discovery. Most of the computational methods developed for predicting DTA use machine learning models, especially deep neural networks, and depend on large-scale labelled data. However, it is difficult to learn enough feature representation from tens of millions of compounds and hundreds of thousands of proteins only based on relatively limited labelled drug-target data. There are a large number of unknown drugs, which never appear in the labelled drug-target data. This is a kind of out-of-distribution problems in bio-medicine. Some recent studies adopted self-supervised pre-training tasks to learn structural information of amino acid sequences for enhancing the feature representation of proteins. However, the task gap between pre-training and DTA prediction brings the catastrophic forgetting problem, which hinders the full application of feature representation in DTA prediction and seriously affects the generalization capability of models for unknown drug discovery. Results To address these problems, we propose the GeneralizedDTA, which is a new DTA prediction model oriented to unknown drug discovery, by combining pre-training and multi-task learning. We introduce self-supervised protein and drug pre-training tasks to learn richer structural information from amino acid sequences of proteins and molecular graphs of drug compounds, in order to alleviate the problem of high variance caused by encoding based on deep neural networks and accelerate the convergence of prediction model on small-scale labelled data. We also develop a multi-task learning framework with a dual adaptation mechanism to narrow the task gap between pre-training and prediction for preventing overfitting and improving the generalization capability of DTA prediction model on unknown drug discovery. To validate the effectiveness of our model, we construct an unknown drug data set to simulate the scenario of unknown drug discovery. Compared with existing DTA prediction models, the experimental results show that our model has the higher generalization capability in the DTA prediction of unknown drugs. Conclusions The advantages of our model are mainly attributed to two kinds of pre-training tasks and the multi-task learning framework, which can learn richer structural information of proteins and drugs from large-scale unlabeled data, and then effectively integrate it into the downstream prediction task for obtaining a high-quality DTA prediction in unknown drug discovery.

show abstract

Section: Resultsmentioning

confidence: 99%

GeneralizedDTA: combining pre-training and multi-task learning to predict drug-target binding affinity for unknown drug discovery

2022

View full text Add to dashboard Cite

show abstract

“…The emergence of AlphaFold fills such a defect to a certain extent, and the present study shows that there is no significant difference in the predictive performance between the two protein structures obtained from AlphaFold2 and experiments. This finding is encouragingas the 3D structure of proteins is of great importance in determining protein functionality, the availability of reliable protein structures will benefit almost all the current deep learning methods for protein-related tasks, including not only protein function prediction approaches (e.g., 1D sequence-based methods ,, and integrated data-based methods , ) but also compound–protein interaction models. , Although 3D structure-based methods have emerged in recent years, the performance of these methods is limited by the number of training samples and is expected to be further improved with the expansion of the data set.…”

Section: Discussionmentioning

confidence: 99%

Enhancing Protein Function Prediction Performance by Utilizing AlphaFold-Predicted Protein Structures

Zhang

et al. 2022

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

The structure of a protein is of great importance in determining its functionality, and this characteristic can be leveraged to train data-driven prediction models. However, the limited number of available protein structures severely limits the performance of these models. AlphaFold2 and its open-source data set of predicted protein structures have provided a promising solution to this problem, and these predicted structures are expected to benefit the model performance by increasing the number of training samples. In this work, we constructed a new data set that acted as a benchmark and implemented a state-of-the-art structure-based approach for determining whether the performance of the function prediction model can be improved by putting additional AlphaFold-predicted structures into the training set and further compared the performance differences between two models separately trained with real structures only and AlphaFold-predicted structures only. Experimental results indicated that structure-based protein function prediction models could benefit from virtual training data consisting of AlphaFold-predicted structures. First, model performances were improved in all three categories of Gene Ontology terms (GO terms) after adding predicted structures as training samples. Second, the model trained only on AlphaFold-predicted virtual samples achieved comparable performances to the model based on experimentally solved real structures, suggesting that predicted structures were almost equally effective in predicting protein functionality.

show abstract

“…The affinity of ligands and proteins is a pivotal parameter for screening potential drug candidate. 26 BatchDTA is an improved platform that successfully alleviates the influence of the batch effects and far overweighs the other methods. It could be done through PaddleHelix platform (https://paddlehelix.baidu.com/app/drug/admet/ train).…”

Section: Ligand-target Affinity Estimation Via Batchdtamentioning

confidence: 99%

The discovery of subunit-selective GluN1/GluN2B NMDAR antagonist via pharmacophere-based virtual screening

Tang,

Jin,

Huang

et al. 2023

Exp Biol Med (Maywood)

View full text Add to dashboard Cite

The incidence and mortality rates of neurodegenerative diseases, such as Alzheimer’s disease and Parkinson’s disease, are gradually increasing worldwide. Numerous studies have demonstrated that N-methyl-D-aspartic acid receptor (NMDAR)-mediated excitotoxicity contributes to neurodegenerative diseases. Ifenprodil, a subtype-selective NMDAR antagonist, showed strong therapeutic potential. However, it suffers from low oral bioavailability and off-target side effects. In this study, natural compounds were identified for selective inhibition of GluN1/GluN2B NMDAR of human. We obtained a set of natural compounds ( n = 81,366) from COCONUT, TIPdb, PAMDB, CMNPD, YMDB, and NPAtlas databases, and then virtually screened by an ifenprodil-structure-based pharmacophore model and molecular docking. The top 100 compounds were selected for binding affinity prediction via batch drug–target affinity (BatchDTA). Then, the top 50 compounds were analyzed by absorption, distribution, metabolism, excretion, toxicity (ADMET). Molecular dynamics involving molecular mechanics/position-Boltzmann surface area (MM-PBSA) calculation were performed to further screening. The top 15 compounds with strong binding affinity and ifenprodil, a proven GluN2B-selective NMDAR antagonist, were subjected to molecular dynamic simulations (100 ns), root-mean-square deviation (RMSD), root-mean-square fluctuation (RMSF), radius of gyration (Rg), H-bonds, solvent accessible surface area (SASA), principal component analysis (PCA), and binding free energy calculations. Based on these analyses, one possible lead compound carrying positive charges (CNP0099440) was identified, with great binding affinity and less off-target activity by contrast to ifenprodil. CNP0099440 has great potential to be a GluN1/GluN2B NMDAR antagonist candidate and can be further detected via in vitro and in vivo experiments.

show abstract

SAG-DTA: Prediction of Drug–Target Affinity Using Self-Attention Graph Network

Cited by 35 publications

References 34 publications

GeneralizedDTA: combining pre-training and multi-task learning to predict drug-target binding affinity for unknown drug discovery

GeneralizedDTA: combining pre-training and multi-task learning to predict drug-target binding affinity for unknown drug discovery

Enhancing Protein Function Prediction Performance by Utilizing AlphaFold-Predicted Protein Structures

The discovery of subunit-selective GluN1/GluN2B NMDAR antagonist via pharmacophere-based virtual screening

Contact Info

Product

Resources

About