基于多模态栈式混合自编码器的药物靶标相互作用预测

doi:10.3778/j.issn.1002-8331.2306-0269

摘要/Abstract

摘要： 针对药物-靶标相互作用（drug-target interactions，DTI）预测中存在的单模态特征不足、多模态数据利用不充分、数据噪声大等问题，提出了一种基于多模态栈式混合自编码器（stacked hybrid autoencoder, SHADTI）的药物靶标相互作用预测方法。SHADTI包括数据预处理模块、潜在特征提取模块和预测模块三部分。数据预处理模块利用随机游走和PPMI（positive pointwise mutual information）算法对药物和靶标的多模态数据进行全局拓扑结构处理。潜在特征提取模块利用深度自编码器混合了降噪块、稀疏块、堆栈块，充分挖掘多模态之间蕴含的信息，生成潜在药物靶标特征向量。预测模块将药物和靶标的潜在特征拼接后输入到全连接层进行预测。所提方法在5个公开数据集上与现有深度学习方法进行对比，实验结果均优于所对比的方法，表明SHADTI能够有效利用多模态数据间的互补信息，提高了DTI预测精度。

关键词: 药物-靶标相互作用, 多模态, 自编码器, 深度学习

Abstract: In order to address the issues of insufficient unimodal features, underutilization of multimodal data, and high data noise in drug-target interaction prediction (DTI), this paper proposes a drug target interaction prediction method based on multimodal stack hybrid autoencoder (SHADTI). SHADTI consists of three components: a data preprocessing module, a latent feature extraction module, and a prediction module. The data preprocessing module employs the random walk with restart (RWR) and positive pointwise mutual information (PPMI) algorithms to process the global topological structure of the multimodal data of drugs and targets. The latent feature extraction module utilizes a deep autoencoder that integrates denoise blocks, sparse blocks, and stacked blocks to effectively mine the information embedded in the multiple modalities, thereby generating latent drug-target feature vectors. The prediction module concatenates the latent features of drugs and targets and feeds them into a fully connected layer for prediction. The proposed method is compared with existing deep learning methods on five publicly available datasets, and the experimental results consistently outperform the compared methods. The results show that SHADTI effectively leverages complementary information from multimodal data and improves the accuracy of DTI prediction.

Key words: drug-target interactions, multimodal, autoencoder, deep learning

张星宇, 陈卓, 黄印, 原雨婷, 李颖, 王彬. 基于多模态栈式混合自编码器的药物靶标相互作用预测[J]. 计算机工程与应用, 2024, 60(19): 334-342.

ZHANG Xingyu, CHEN Zhuo, HUANG Yin, YUAN Yuting, LI Ying, WANG Bin. Prediction of Drug-Target Interactions Based on Multimodal Stacked Hybrid Autoencoder[J]. Computer Engineering and Applications, 2024, 60(19): 334-342.

参考文献

[1] PENG J, WANG Y, GUAN J, et al. An end-to-end heterogeneous graph representation learning-based framework for drug-target interaction prediction[J]. Briefings in Bioinformatics, 2021, 22(5): bbaa430.
[2] XUAN P, FAN M, CUI H, et al. GVDTI: graph convolu-tional and variational autoencoders with attribute-level at-tention for drug-protein interaction prediction[J]. Briefings in Bioinformatics, 2022, 23(1): bbab453.
[3] SHAO K, ZHANG Y, WEN Y, et al. DTI-HETA: prediction of drug-target interactions based on GCN and GAT on heterogeneous graph[J]. Briefings in Bioinformatics, 2022, 23(3): bbac109.
[4] 李卓容, 唐云祁. 基于深度学习的多模态生物特征融合模型[J]. 计算机工程与应用, 2023, 59(7): 180-189.
LI Z R, TANG Y Q. Multimodal biometric fusion model based on deep learning[J]. Computer Engineering and Ap-plications, 2023, 59(7): 180-189.
[5] ZHANG P, WEI Z, CHE C, et al. DeepMGT-DTI: transformer network incorporating multilayer graph information for drug-target interaction prediction[J]. Computers in Biology and Medicine, 2022, 142: 105214.
[6] BAI P, MILJKOVI? F, JOHN B, et al. Interpretable bilinear attention network with domain adaptation improves drug-target prediction[J]. Nature Machine Intelligence, 2023, 5(2): 126-136.
[7] WU Y, GAO M, ZENG M, et al. BridgeDPI: a novel graph neural network for predicting drug-protein interactions[J]. Bioinformatics, 2022, 38(9): 2571-2578.
[8] HUA Y, SONG X, FENG Z, et al. MFR-DTA: a multi-functional and robust model for predicting drug-target binding affinity and region[J]. Bioinformatics, 2023, 39(2): btad056.
[9] 谢良旭, 李峰, 谢建平, 等. 基于融合神经网络模型的药物分子性质预测[J]. 计算机科学, 2021, 48(9): 251-256.
XIE L X, LI F, XIE J P, et al. Predicting drug molecular properties based on ensembling neural networks models[J]. Computer Science, 2021, 48(9): 251-256.
[10] WANG H, WANG J, DONG C, et al. A novel approach for drug-target interactions prediction based on multimodal deep autoencoder[J]. Frontiers in Pharmacology, 2020, 10: 1592.
[11] LIU Z, CHEN Q, LAN W, et al. GADTI: graph autoencoder approach for DTI prediction from heterogeneous network[J]. Frontiers in Genetics, 2021, 12: 650821.
[12] SAJADI S Z, ZARE CHAHOOKI M A, GHARAGHANI S, et al. AutoDTI++: deep unsupervised learning for DTI prediction by autoencoders[J]. BMC Bioinformatics, 2021, 22(1): 1-19.
[13] 曹业伟, 刘飞. 癌症多组学数据深度自编码器整合分型方法[J]. 计算机工程与应用, 2022, 58(18): 154-161.
CAO Y W, LIU F. Multi-omics data deep autoencoder in-tegration for cancer subtyping[J]. Computer Engineering and Applications, 2022, 58(18): 154-161.
[14] CHU Y, KAUSHIK A C, WANG X, et al. DTI-CDF: a cascade deep forest model towards the prediction of drug-target interactions based on hybrid features[J]. Briefings in Bioinformatics, 2021, 22(1): 451-462.
[15] OLAYAN R S, ASHOOR H, BAJIC V B. DDR: efficient computational method to predict drug-target interactions using graph mining and machine learning approaches[J]. Bioinformatics, 2018, 34(7): 1164-1173.
[16] HUANG K, XIAO C, GLASS L M, et al. MolTrans: mo-lecular interaction transformer for drug-target interaction prediction[J]. Bioinformatics, 2021, 37(6): 830-836.
[17] KANEHISA M, FURUMICHI M, SATO Y, et al. KEGG: integrating viruses and cellular organisms[J]. Nucleic Acids Research, 2021, 49(D1): D545-D551.
[18] WANG X, CHENG Y, YANG Y, et al. Multitask joint strategies of self-supervised representation learning on biomedical networks for drug discovery[J]. Nature Machine Intelligence, 2023, 5(4): 445-456.
[19] WAN F, HONG L, XIAO A, et al. NeoDTI: neural integra-tion of neighbor information from a heterogeneous network for discovering new drug-target interactions[J]. Bio-informatics, 2019, 35(1): 104-111.
[20] WEN J, ZHANG X, RUSH E, et al. Multimodal represen-tation learning for predicting molecule-disease relations[J]. Bioinformatics, 2023, 39(2): btad085.
[21] CHATTERJEE A, WALTERS R, SHAFI Z, et al. Improving the generalizability of protein-ligand binding predictions with AI-Bind[J]. Nature Communications, 2023, 14(1): 1989.
[22] FAN X N, ZHANG S W, ZHANG S Y, et al. Prediction of lncRNA-disease associations by integrating diverse heter-ogeneous information sources with RWR algorithm and positive pointwise mutual information[J]. BMC Bioinfor-matics, 2019, 20(1): 1-12.
[23] CHANG J W, DING Y, TAHIR UL QAMAR M, et al. A deep learning model based on sparse auto-encoder for pri-oritizing cancer-related genes and drug target combinations[J]. Carcinogenesis, 2019, 40(5): 624-632.
[24] MENDEZ D, GAULTON A, BENTO A P, et al. ChEMBL: towards direct deposition of bioassay data[J]. Nucleic Acids Research, 2019, 47(D1): D930-D940.
[25] WISHART D S, FEUNANG Y D, GUO A C, et al. DrugBank 5.0: a major update to the DrugBank database for 2018[J]. Nucleic Acids Research, 2018, 46(D1): D1074-D1082.
[26] WANG Y, ZHANG S, LI F, et al. Therapeutic target database 2020: enriched resource for facilitating research and early development of targeted therapeutics[J]. Nucleic Acids Research, 2020, 48(D1): D1031-D1041.
[27] KIM S, CHEN J, CHENG T, et al. PubChem in 2021: new data content and improved web interfaces[J]. Nucleic Acids Research, 2021, 49(D1): D1388-D1395.