TOP: A deep mixture representation learning method for boosting molecular toxicity prediction

Peng, Yuzhong; Zhang, Ziqiao; Jiang, Qizhi; Guan, Jihong; Zhou, Shuigeng

doi:10.1016/j.ymeth.2020.05.013

Cited by 26 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the past decade, the rapid development and wide application of artificial intelligence (AI) techniques have shown its great success in many areas, especially in computer vision and natural language processing (NLP). With the increasing amassment of accessible drug data, AI techniques are being introduced into drug discovery, and a number of AI-based models for molecular property prediction have been developed ( Jiménez et al , 2018 ; Liew et al , 2009 ; Melville et al , 2009 ; Peng et al , 2020 ). Particularly, with the development of graph neural networks (GNNs) in recent years ( Kipf and Welling, 2017 ; Veličković et al , 2018 ), graph-based molecular property prediction is becoming a hot research topic ( Coley et al , 2017 ; Duvenaud et al , 2015 ; Gilmer et al , 2017 ; Kearnes et al , 2016 ; Xiong et al , 2020 ; Zhou and Li, 2017 ).…”

Section: Introductionmentioning

confidence: 99%

FraGAT: a fragment-oriented multi-scale graph attention model for molecular property prediction

2021

Self Cite

View full text Add to dashboard Cite

Motivation Molecular property prediction is a hot topic in recent years. Existing graph-based models ignore the hierarchical structures of molecules. According to the knowledge of chemistry and pharmacy, the functional groups of molecules are closely related to its physio-chemical properties and binding affinities. So, it should be helpful to represent molecular graphs by fragments that contain functional groups for molecular property prediction. Results In this paper, to boost the performance of molecule property prediction, we first propose a definition of molecule graph fragments that may be or contain functional groups, which are relevant to molecular properties, then develop a fragment-oriented multi-scale graph attention network for molecular property prediction, which is called FraGAT. Experiments on several widely-used benchmarks are conducted to evaluate FraGAT. Experimental results show that FraGAT achieves state-of-the-art predictive performance in most cases. Furthermore, our case studies showthat when the fragments used to represent the molecule graphs contain functional groups, the model can make better predictions. This conforms to our expectation and demonstrates the interpretability of the proposed model. Availability and implementation The code and data underlying this work are available in GitHub, at https://github.com/ZiqiaoZhang/FraGAT. Supplementary information Supplementary data are available at Bioinformatics online.

show abstract

Section: Introductionmentioning

confidence: 99%

FraGAT: a fragment-oriented multi-scale graph attention model for molecular property prediction

2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…RNN based methods have been widely used for tasks in absorption category and Tox21 datasets (Table S1). TOP [112] achieved an excellent performance in toxicity prediction using Tox21 dataset (mean AUC-ROC: 0.950) by integrating shallow representation on SMILES into biGRU in combination with some molecular descriptors (logP, MW and TPSA). By incorporating the physiochemical properties, TOP resulted in 0.195 performance gain in terms of AUC-ROC.…”

Section: Deep Learning Technologies: How Well Can We Accomplish the T...mentioning

confidence: 99%

“…logP (partition coefficient) and TPSA (total polar surface area) are related to the solubility of a compound in aqueous solutions and the presence of specific structural features such as the number of rings is related to carcinogenesis [205] . TOP [112] leveraged logP, molecular weight, and TPSA to the independent fully connected layers on the word embeddings of SMILES along with biGRU to learn chemical structures. Addition of physico-chemical properties selected by genetic algorithm to biGRU featurization provided 0.195 of improvement in AUC-ROC to toxicity prediction tasks.…”

Section: Additional Features Required Beyond Chemical Compound Inform...mentioning

confidence: 99%

On modeling and utilizing chemical compound information with deep learning technologies: A task-oriented approach

Lim¹,

Lee²,

Piao³

et al. 2022

Computational and Structural Biotechnology Journal

View full text Add to dashboard Cite

“…The most commonly used representations of drugs and targets are one-dimensional (1D) sequences such as Simplified Molecular Input Line Entry Specification (SMILES) strings for drugs and amino acid sequences for targets ( Karimi et al , 2019 ; Liu et al , 2020 ; Peng et al , 2020 ; Tsubaki et al , 2019 ; Zheng et al , 2020 ). The models with 1D sequences as input generally use Convolutional Neural network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM) blocks or Gated Recurrent Unit (GRU) blocks to extract drug and target features.…”

Section: Introductionmentioning

confidence: 99%

Effective drug–target interaction prediction with mutual interaction neural network

Zhang

Guan

et al. 2022

Bioinformatics

Self Cite

View full text Add to dashboard Cite

Motivation Accurately predicting drug-target interaction (DTI) is a crucial step to drug discovery. Recently, deep learning techniques have been widely used for DTI prediction and achieved significant performance improvement. One challenge in building deep learning models for DTI prediction is how to appropriately represent drugs and targets. Target distance map and molecular graph are low dimensional and informative representations, which however have not been jointly used in DTI prediction. Another challenge is how to effectively model the mutual impact between drugs and targets. Though attention mechanism has been employed to capture the one-way impact of targets on drugs or vice versa, the mutual impact between drugs and targets has not yet been explored, which is very important in predicting their interactions. Results Therefore, in this paper we propose MINN-DTI, a new model for DTI prediction. MINN-DTI combines an interacting-transformer module (called Interformer) with an improved Communicative Message Passing Neural Network (CMPNN) (called Inter-CMPNN) to better capture the two-way impact between drugs and targets, which are represented by molecular graph and distance map respectively. The proposed method obtains better performance than the state-of-the-art methods on three benchmark datasets: DUD-E, human and BindingDB. MINN-DTI also provides good interpretability by assigning larger weights to the amino acids and atoms that contribute more to the interactions between drugs and targets. Availability The data and code of this study are available at https://github.com/admislf/MINN-DTI.

show abstract

TOP: A deep mixture representation learning method for boosting molecular toxicity prediction

Cited by 26 publications

References 27 publications

FraGAT: a fragment-oriented multi-scale graph attention model for molecular property prediction

FraGAT: a fragment-oriented multi-scale graph attention model for molecular property prediction

On modeling and utilizing chemical compound information with deep learning technologies: A task-oriented approach

Effective drug–target interaction prediction with mutual interaction neural network

Contact Info

Product

Resources

About