Article
Biochemical Research Methods
Ali Akbar Jamali, Anthony Kusalik, Fang-Xiang Wu
Summary: Prediction of drug-target interactions (DTIs) is important in drug development and discovery. Computational DTI prediction is a cost-effective shortcut compared to experimental methods. This study proposes an effective approach, NMTF-DTI, which utilizes multiple kernels and Laplacian regularization to improve prediction performance.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2023)
Article
Biochemical Research Methods
Bin Liu, Dimitrios Papadopoulos, Fragkiskos D. Malliaros, Grigorios Tsoumakas, Apostolos N. Papadopoulos
Summary: The study proposes a novel optimization framework called Multiple similarity DeepWalk-based Matrix Factorization (MDMF) to predict drug-target interactions, achieving unified embedding generation and interaction prediction, as well as improved accuracy and novelty in prediction.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Multidisciplinary Sciences
Kohei Fukuto, Tatsuya Takagi, Yu-Shi Tian
Summary: This study constructed a more accurate side effect prediction model using the logistic matrix factorization algorithm, addressing importance differences and cold-start problems through a weighting strategy and mapping method, improving prediction accuracy and potential applications in clinical warning systems.
SCIENTIFIC REPORTS
(2021)
Article
Computer Science, Information Systems
Stuti Jain, Emilie Chouzenoux, Kriti Kumar, Angshul Majumdar
Summary: Simultaneous co-administration of multiple drugs may lead to adverse drug reactions. Identifying drug-drug interactions (DDIs) is crucial for drug development and repurposing of old drugs. This paper introduces a novel Graph Regularized Probabilistic Matrix Factorization (GRPMF) method that incorporates expert knowledge through graph-based regularization within a matrix factorization framework. An efficient optimization algorithm is proposed to solve the resulting non-convex problem. Evaluation using the DrugBank dataset demonstrates the superior performance of GRPMF compared to other techniques.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Computer Science, Interdisciplinary Applications
Liyi Yu, Meiling Cheng, Wangren Qiu, Xuan Xiao, Weizhong Lin
Summary: Unexpected side effects are a major factor leading to the failure of drug trials, so discovering potential side effects in a computational manner can improve the success rate of drug screening. However, previous methods have mainly focused on a single perspective, neglecting the need to consider multiple types of information about drugs. In this study, a hybrid embedding graph neural network model is proposed to integrate various drug features and predict potential side effects, outperforming other advanced methods.
JOURNAL OF BIOMEDICAL INFORMATICS
(2022)
Article
Biochemical Research Methods
Maryam Bagherian, Renaid B. Kim, Cheng Jiang, Maureen A. Sartor, Harm Derksen, Kayvan Najarian
Summary: Predicting interactions between drugs and targets is crucial in new drug discovery and drug repurposing. Developing efficient prediction methods is essential to avoid costly experimental processes. Matrix factorization methods have been proven to be reliable, and the proposed CMMC and CTMC outperform other methods in terms of performance and runtime.
BRIEFINGS IN BIOINFORMATICS
(2021)
Review
Biochemical Research Methods
Jinyu Chen, Louxin Zhang
Summary: This study systematically evaluates 17 representative methods for drug response prediction developed in the past 5 years on four large public datasets with nine metrics, providing insights and lessons for future research into drug response prediction.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Biochemical Research Methods
Junjun Zhang, Minzhu Xie
Summary: In this study, a new method (ADA-GRMFC) is proposed to predict drug-target interactions by constructing similarity matrices for drugs and targets, and using non-negative matrix factorization combined with prior knowledge constraints. Experimental results show that this method performs better than other existing methods.
BMC BIOINFORMATICS
(2022)
Article
Computer Science, Hardware & Architecture
Li-Gang Gao, Meng-Yun Yang, Jian-Xin Wang
Summary: In this study, a novel matrix factorization method called collaborative matrix factorization with soft regularization (SRCMF) is proposed to improve prediction performance of drug-target interactions by combining drug and target similarity information with matrix factorization. SRCMF aims to approximate similarity features and potential features of DTI, rather than making them identical. Experimental results show that SRCMF achieves better prediction results in terms of AUPR compared to six state-of-the-art methods, and outperforms collaborative matrix factorization under different noise levels of similarity data.
JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY
(2021)
Article
Chemistry, Medicinal
Iori Azuma, Tadahaya Mizuno, Hiroyuki Kusuhara
Summary: This study proposed a new method called neighborhood regularized bidirectional matrix factorization (NRBdMF) to predict drug effects by incorporating bidirectionality. The NRBdMF model achieved high accuracy and interpretability in predicting both side effects and therapeutic effects.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2023)
Article
Biochemical Research Methods
Raziyeh Masumshah, Rosa Aghdam, Changiz Eslahchi
Summary: This study presents a neural network-based method for predicting polypharmacy side effects, which outperforms five well-established methods in terms of accuracy, complexity, and running time speed. The proposed method utilizes novel feature vectors and drug-protein interaction information to improve prediction efficiency.
BMC BIOINFORMATICS
(2021)
Article
Biochemical Research Methods
Yijie Ding, Hongmei Zhou, Quan Zou, Lei Yuan
Summary: This article introduces the types and risks of adverse drug reactions, emphasizes the importance of monitoring drug side effects, and proposes a machine learning method for predicting drug side effects. The method, called CLMF-NTK, achieves the best predictive performance compared to other computational methods according to the test results.
Article
Biochemical Research Methods
S. Morteza Hashemi, Arash Zabihian, Mohsen Hooshmand, Sajjad Gharaghani
Summary: Due to the high resource consumption of introducing a new drug, drug repurposing is important in drug discovery. Matrix factorization methods have drawbacks in drug-target interaction (DTI) prediction, and thus a deep learning model (DRaW) is proposed to overcome these issues and perform better than other methods on COVID-19 datasets.
BMC BIOINFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Chun Yen Lee, Yi-Ping Phoebe Chen
Summary: In this study, a hybrid deep learning model utilizing graph convolutional neural network and bidirectional long short-term memory recurrent neural networks was developed to predict drug side-effects, achieving significant achievements in accuracy scores.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2021)
Article
Psychiatry
Yongbiao Zhao, Yuanyuan Ma, Qilin Zhang
Summary: In this study, a novel method called logical matrix factorization and local nearest neighbor constraints (LMFLNC) was proposed for predicting metabolite-disease interactions. By taking into account the local tiny structure of metabolites and diseases in the similarity networks, the LMFLNC method achieved higher accuracy in interaction prediction.
FRONTIERS IN PSYCHIATRY
(2023)
Article
Biochemical Research Methods
Zun Liu, Xing-Nan Wang, Hui Yu, Jian-Yu Shi, Wen-Min Dong
Summary: This paper proposes a cold start prediction model CSMDDI for both single-type and multiple-type drug-drug interactions. CSMDDI can predict not only whether pharmacological reactions occur between two drugs, but also what reaction types they induce in the cold start scenario. The comparison with state-of-the-art methods demonstrates that CSMDDI achieves good performance in DDI prediction in the cold start scenario, both in terms of occurrence prediction and multi-type reaction prediction.
BMC BIOINFORMATICS
(2022)
Article
Biochemical Research Methods
Arnold K. Nyamabo, Hui Yu, Zun Liu, Jian-Yu Shi
Summary: Drug-drug interactions (DDIs) can have adverse effects on the body. Researchers introduce a method called GMPNN-CS, which uses a gated message passing neural network (GMPNN) to predict interactions between pairs of drugs. By learning chemical substructures from molecular graph representations, this method provides accurate predictions.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Computer Science, Information Systems
Hui Yu, WenMin Dong, JianYu Shi
Summary: A relation-aware network embedding model RANEDDI is proposed for drug-drug interaction prediction, showing superior performance in binary and multirelational DDIs. By integrating multirelational embedding and relation-aware network structure embedding of drugs, RANEDDI effectively improves DDI prediction accuracy and demonstrates robustness in binary DDI prediction, even with limited labeled data. Available source code at https://github.com/DongWenMin/RANEDDI.
INFORMATION SCIENCES
(2022)
Article
Medicine, Research & Experimental
Pengcheng Zhao, Lin Lin, Mozheng Wu, Lili Wang, Qi Geng, Li Li, Ning Zhao, Jianyu Shi, Cheng Lu
Summary: This study proposes a novel method based on Cosine-correlation and Similarity-comparison of Local Network (CSLN) for preliminary screening and assignment of weights to targets of fresh natural drug molecules. The results show that CSLN outperforms existing prediction models and can be an alternative strategy for screening targets of fresh natural drug molecules.
JOURNAL OF TRANSLATIONAL MEDICINE
(2022)
Article
Biochemical Research Methods
Yue-Hua Feng, Shao-Wu Zhang, Qing-Qing Zhang, Chu-Han Zhang, Jian-Yu Shi
Summary: A novel deep learning-based method, deepMDDI, was proposed in the study to predict drug-drug interactions efficiently and accurately. Experimental results demonstrated its superiority over other state-of-the-art deep learning methods and its capability to discover new valid DDIs.
ANALYTICAL BIOCHEMISTRY
(2022)
Article
Biochemical Research Methods
Yue-Hua Feng, Shao-Wu Zhang, Yi-Yang Feng, Qing-Qing Zhang, Ming-Hui Shi, Jian-Yu Shi
Summary: Current machine learning-based methods have made impressive predictions in mono-type and multi-type drug-drug interactions (DDIs) scenarios, but they fail to consider the enhancing and depressive pharmacological changes triggered by DDIs. Additionally, these pharmacological changes are asymmetric due to the different roles of the two drugs involved in an interaction. To address these issues, this study utilizes Balance and Status theories from social networks to reveal the topological patterns in directed pharmacological DDIs. A novel graph representation learning model named SGRL-DDI is then proposed to achieve multitask prediction of DDIs by integrating relation graph convolutional networks with Balance and Status patterns. Experimental results using DDI entries collected from DrugBank demonstrate the superiority of the SGRL-DDI model compared to other state-of-the-art methods. Moreover, the practical effectiveness of the model is demonstrated through a version-dependent test.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Automation & Control Systems
Ke Wang, Zicong Chen, Mingjia Zhu, Siu-Ming Yiu, Chien-Ming Chen, Mohammad Mehedi Hassan, Stefano Izzo, Giancario Fortino
Summary: Artificial intelligence-driven automation is becoming the technical trend in the new automation era. Convolutional neural network (CNN) technology has been widely used in industrial automation for defect detection and machine vision-driven automation for robot-human tracking. However, the high dependence on neural networks can lead to potential failures in defect detection system.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Biochemical Research Methods
Hui Yu, KangKang Li, WenMin Dong, ShuangHong Song, Chen Gao, JianYu Shi
Summary: Drug-drug interactions (DDI) can cause adverse reactions in the body, and accurate prediction of DDI can reduce medical risks. Existing prediction methods mainly focus on drug-related features or DDI networks, neglecting the potential information in drug-related biological entities. To overcome this limitation, we propose an attention-based cross domain graph neural network (ACDGNN) that considers different drug-related entities and uses cross-domain operation to propagate information. ACDGNN demonstrates superior performance in predicting DDIs compared to state-of-the-art models.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Biochemical Research Methods
Hui Yu, KangKang Li, JianYu Shi
Summary: To accurately predict drug-drug interaction (DDI) events, we propose a deep learning model named DGANDDI, which utilizes an adversarial learning strategy. DGANDDI incorporates drug attribute and topological information of DDI network using a two-GAN architecture, enabling more comprehensive drug representations. Experimental results demonstrate that DGANDDI effectively predicts DDI occurrence and outperforms state-of-the-art models.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Hui Yu, Qiao Feng Wang, Jian Yu Shi
Summary: In the field of data mining, the performance of clustering is heavily influenced by the number of samples. However, obtaining enough data samples can be difficult and expensive in some applications. To address this problem, a new data augmentation method called GMM_WGAN is proposed, which combines the Gaussian Mixture Model and Wasserstein generative adversarial network. The method captures the potential distribution of the real dataset and generates additional data samples to augment the small size dataset. Experimental results demonstrate that GMM_WGAN outperforms other data enhancement methods based on five evaluation metrics in clustering 10 small size datasets.
NEURAL PROCESSING LETTERS
(2023)
Article
Computer Science, Information Systems
Yujing Sun, Hao Xiong, Siu Ming Yiu, Kwok Yan Lam
Summary: Bitcoin is gaining popularity but lacks tools for effective investigation of transactions. To address this, we present BitAnalysis, a novel visualization system for interactive bitcoin wallet investigation. BitAnalysis was developed based on the advice and requirements of bitcoin-related entrepreneurs and regulators. It provides law-enforcement officers and regulators with intuitive visual interfaces and functions to analyze and track bitcoin transactions, identify wallet correlation, and visualize the flow of bitcoins. We also conducted a user study to validate the effectiveness and usability of BitAnalysis.
IEEE TRANSACTIONS ON BIG DATA
(2023)
Review
Biochemical Research Methods
Xuan Lin, Lichang Dai, Yafang Zhou, Zu-Guo Yu, Wen Zhang, Jian-Yu Shi, Dong-Sheng Cao, Li Zeng, Haowen Chen, Bosheng Song, Philip S. Yu, Xiangxiang Zeng
Summary: Recent advances in AI and deep learning models have proven their usefulness in biomedical applications, specifically in predicting drug-drug interactions (DDIs). Traditional clinical trials and experiments for DDIs prediction are time-consuming and expensive. The application of AI and deep learning in this field faces challenges such as data availability and encoding, as well as computational method design. This review summarizes various methods for DDIs prediction and provides a comprehensive guide for researchers and developers.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Qing-Qing Zhang, Shao-Wu Zhang, Yue-Hua Feng, Jian-Yu Shi
Summary: A novel method called HyperSynergy is proposed to address the drug synergy prediction problem in data-poor cell lines. It utilizes a prior-guided Hypernetwork architecture to generate cell line-dependent parameters for drug synergy prediction. Experimental results demonstrate that HyperSynergy outperforms other methods not only on data-poor cell lines but also on data-rich cell lines.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Hui Yu, Jing Wang, Shi-Yu Zhao, Omayo Silver, Zun Liu, Jingtao Yao, Jian-Yu Shi
Summary: Deep learning-based models have limited interpretability in predicting drug-drug interactions (DDIs). We propose a novel approach that uses granular computing to identify key substructures and achieves high accuracy in predicting DDIs.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Computer Science, Theory & Methods
Zhiwei Wang, Feng Liu, Siuming Yiu, Longwen Lan
Summary: This paper proposes an efficient identity-based fuzzy message detection method that generates a single flag ciphertext quickly through offline and online computations, and applies this method to blockchain technology. Additionally, a history indexing scheme is designed to ensure the order of transaction history and accessing storage cloud, with privacy guarantees and analysis of differential privacy requirements.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2023)