Article
Computer Science, Information Systems
Yao Fu, Runtao Yang, Lina Zhang, Xu Fu
Summary: This paper proposes a novel method for predicting circRNA-disease associations. It constructs a heterogeneous graph network and uses a path embedding model to obtain initial feature vectors, then applies the CosMulformer model for interaction vector acquisition and prediction. Experimental results demonstrate the effectiveness of the method, with case studies validating its applicability to breast cancer and colorectal cancer.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2023)
Article
Biotechnology & Applied Microbiology
Yuwei Guo, Ming Yi
Summary: Circular RNAs (circRNAs) with closed circular structure have been found to play a significant role in reducing diseases and can be potential biomarkers. However, traditional experimental methods are time-consuming. To address this, researchers proposed a novel method called THGNCDA, which uses a graph neural network with attention to predict the association between circRNAs and diseases, considering the information of miRNAs and attributes of circRNA-disease pairs.
BRIEFINGS IN FUNCTIONAL GENOMICS
(2023)
Article
Cell Biology
Wei Peng, Jielin Du, Wei Dai, Wei Lan
Summary: A novel method called MDN-NMTF is proposed for predicting miRNA-disease associations by integrating diverse biological information and considering module properties. The results show that this method significantly improves the prediction of miRNA-disease associations compared to existing methods.
FRONTIERS IN CELL AND DEVELOPMENTAL BIOLOGY
(2021)
Article
Biochemical Research Methods
Abdur Rahman M. A. Basher, Steven J. Hallam
Summary: The pathway2vec software package automatically generates features for pathway inference by building a three-layered network consisting of compounds, enzymes, and pathways. The benchmark results show improved prediction outcomes in metabolic pathway prediction tasks by leveraging embeddings.
Article
Computer Science, Artificial Intelligence
Anping Zhao, Yu Yu
Summary: The proposed method effectively predicts sentiment links among users by combining global structural information with multi-dimensional relations and heterogeneous context information. Experimental results demonstrate the effectiveness of incorporating social relations and profile context information into sentiment link prediction, especially in cold-start scenarios.
COGNITIVE COMPUTATION
(2022)
Article
Biochemical Research Methods
Nansu Zong, Rachael Sze Nga Wong, Yue Yu, Andrew Wen, Ming Huang, Ning Li
Summary: Researchers conducted a study on modularization for network-based prediction, comparing the performance of various methods through experiments and benchmarking to identify the best prediction strategy from a network and methodological perspective. The results showed that the proposed method outperforms existing methods, demonstrating reliability and scalability in disease-specific prediction tasks.
BRIEFINGS IN BIOINFORMATICS
(2021)
Article
Biochemistry & Molecular Biology
Dong-Ling Yu, Zu-Guo Yu, Guo-Sheng Han, Jinyan Li, Vo Anh
Summary: Abnormal miRNA functions play a significant role in various diseases, with up to five different association types identified. Predicting multiple association types is crucial for understanding disease mechanisms, and current methods lack exploiting nonlinear characteristics in the miRNA-disease association network to improve performance.
Article
Biochemical Research Methods
Ming He, Chen Huang, Bo Liu, Yadong Wang, Junyi Li
Summary: This study proposes a new heterogeneous network embedding method FactorHNE for predicting disease-gene association, which optimizes the model through multiple semantic relationships and shows better performance and scalability in experiments.
BMC BIOINFORMATICS
(2021)
Article
Biochemical Research Methods
Xiaosa Zhao, Xiaowei Zhao, Minghao Yin
Summary: In this paper, a novel heterogeneous graph attention network framework (HGATLDA) based on meta-paths is proposed for predicting lncRNA-disease associations. The framework incorporates node features and relationships in the network, and utilizes attention mechanism and neural inductive matrix completion to capture complex associations between lncRNAs and diseases. Experimental results demonstrate the effectiveness of the proposed framework.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Computer Science, Information Systems
Xian Mo, Rui Tang, Hao Liu
Summary: This paper proposes a relation-aware Heterogeneous Graph Convolutional Network architecture for predicting temporal heterogeneous network relationships. It utilizes a continuous-time temporal heterogeneous neighbor generation algorithm to capture continuous-time interactions and learns the most relevant relationship using a relation-aware Heterogeneous Graph Convolutional Network. Experimental results show significant advances in prediction accuracy and efficiency compared to state-of-art approaches.
INFORMATION SCIENCES
(2023)
Article
Genetics & Heredity
Qingyu Liu, Junjie Yu, Yanning Cai, Guishan Zhang, Xianhua Dai
Summary: The study introduces a novel method named SAAED for semantic association analysis of circRNA-disease, achieving the best overall performance compared to other state-of-the-art models and showing promise in predicting potential circRNA-disease associations efficiently.
FRONTIERS IN GENETICS
(2022)
Article
Biochemical Research Methods
Bailong Liu, Xiaoyan Zhu, Lei Zhang, Zhizheng Liang, Zhengwei Li
Summary: The study confirms the important role of miRNAs in diagnosing and treating diseases. A novel combined embedding model is proposed to predict miRNA-disease associations, showing superior performance compared to other methods.
BMC BIOINFORMATICS
(2021)
Article
Genetics & Heredity
Zhixian Liu, Qingfeng Chen, Wei Lan, Haiming Pan, Xinkun Hao, Shirui Pan
Summary: The article introduces a graph autoencoder approach for DTI prediction, which integrates diverse datasets related to drugs and targets. This method outperforms baseline methods in predictive accuracy.
FRONTIERS IN GENETICS
(2021)
Article
Computer Science, Information Systems
Chenji Huang, Yixiang Fang, Xuemin Lin, Xin Cao, Wenjie Zhang
Summary: In this article, we propose a novel prediction model called ABLE, which utilizes the Attention mechanism and BiLSTM for Embedding, to improve the performance of meta-path prediction in heterogeneous information networks. Experimental results show that ABLE outperforms existing methods on multiple real datasets.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2022)
Article
Biochemical Research Methods
Xiaosa Zhao, Jun Wu, Xiaowei Zhao, Minghao Yin
Summary: This study proposes a new multi-view contrastive heterogeneous graph attention network (GAT) method for predicting lncRNA-disease associations. The method constructs two view graphs using rich biological data sources and designs a cross-contrastive learning task to guide graph embeddings. Experimental results show the effectiveness of this method.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Computer Science, Information Systems
Chunlei Tang, Joseph M. Plasek, Xiao Shi, Meihan Wan, Haohan Zhang, Min-Jeoung Kang, Liqin Wang, Sevan M. Dulgarian, Yun Xiong, Jing Ma, David W. Bates, Li Zhou
Summary: This study predicts mortality risk in patients with chronic obstructive pulmonary disease using clinical notes, optimizing the accuracy of linear regression and support vector machines by determining a tolerance range. The results demonstrate an overall improvement in machine learning approaches after considering the optimal tolerance range.
IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
(2021)
Article
Biochemical Research Methods
J. Harry Caufield, John Fu, Ding Wang, Vladimir Guevara-Gonzalez, Wei Wang, Peipei Ping
Summary: Proteomics aims to study protein features in entire systems, with various resources available to make results more discoverable, accessible, interoperable, and reusable. Linking specific terms, identifiers, and texts can unify individual data points, potentially revealing new relationships and maximizing the value of datasets and methods for the proteomics community and beyond.
JOURNAL OF PROTEOME RESEARCH
(2021)
Article
Computer Science, Theory & Methods
Youfu Li, Matteo Interlandi, Fotis Psallidas, Wei Wang, Carlo Zaniolo
Summary: Many DISC systems provide easy-to-use APIs and efficient scheduling and execution strategies for building concise data-parallel programs. However, some crucial features and optimizations are not well-supported, requiring runtime dataflow states to achieve.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2021)
Article
Biochemical Research Methods
Jyun-Yu Jiang, Chelsea J-T Ju, Junheng Hao, Muhao Chen, Wei Wang
Summary: circRNA is a novel class of long non-coding RNAs that play important roles in gene regulation and disease association. The JEDI framework, utilizing deep learning and a cross-attention layer, effectively predicts circRNAs, outperforming existing methods significantly.
Editorial Material
Health Care Sciences & Services
Chunlei Tang, Joseph M. Plasek, Suhua Zhang, Yun Xiong, Yangyong Zhu, Jing Ma, L. Zhou, David W. Bates
Summary: Big data epidemiology provides data-driven insights for pandemic response, utilizing tools different from traditional methods. Addressing issues like insufficient data and data inaccessibility requires combining techniques across disciplines.
INTERNATIONAL JOURNAL FOR QUALITY IN HEALTH CARE
(2021)
Article
Multidisciplinary Sciences
Jyun-Yu Jiang, Yichao Zhou, Xiusi Chen, Yan-Ru Jhou, Liqi Zhao, Sabrina Liu, Po-Chun Yang, Jule Ahmar, Wei Wang
Summary: This paper proposes a method to leverage social media users as social sensors, predicting pandemic trends while suggesting potential risk factors for public health experts. The method utilizes deep learning models to recognize important entities and their relations, establishing dynamic heterogeneous graphs to describe the observations of social media users. A web-based system is also developed to allow easy interaction for domain experts without computer science backgrounds.
PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES
(2022)
Article
Computer Science, Information Systems
Justin Wood, Corey Arnold, Wei Wang
Summary: Recent work suggests incorporating knowledge sources into the topic modeling process to improve topic discovery. However, existing semi-supervised topic models assume that the corpus contains topics on a subset of a domain, leading to slow inference when considering a large number of article-topics. This paper presents a ranking technique based on the PageRank algorithm to speed up the inference process and improve perplexity and interpretability. The results show significant improvements in various evaluation metrics compared to baseline methods.
Article
Health Care Sciences & Services
Chunlei Tang, Jing Ma, Li Zhou, Joseph Plasek, Yuqing He, Yun Xiong, Yangyong Zhu, Yajun Huang, David Bates
Summary: Organizational, administrative, and educational challenges hinder the efficient utilization of Research Patient Data Repositories (RPDRs) in biomedical data science infrastructures. This article explores applying data science thinking and practices from the business sector, known as the data industry viewpoint, to enhance RPDRs.
JOURNAL OF MEDICAL INTERNET RESEARCH
(2022)
Article
Computer Science, Information Systems
Yucheng Jin, Yun Xiong, Dan Shi, Yifei Lin, Lifang He, Yao Zhang, Joseph M. Plasek, Li Zhou, David W. Bates, Chunlei Tang
Summary: This study aims to develop an objective and unbiased method for learning automatic coding algorithms from clinical records with partial relevant codes. By using positive-unlabeled learning with reweighting and integrating supervision from an annotation tool, the performance of the algorithms is significantly improved, addressing the issues of annotation noise and imbalance.
JOURNAL OF THE AMERICAN MEDICAL INFORMATICS ASSOCIATION
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Seungbae Kim, Jyun-Yu Jiang, Wei Wang
Summary: In this study, the SPoD model is proposed to detect undisclosed sponsorship in social media posts by learning various aspects of the posts. The experimental results demonstrate that SPoD significantly out-performs existing baseline methods in discovering sponsored posts on social media.
WSDM '21: PROCEEDINGS OF THE 14TH ACM INTERNATIONAL CONFERENCE ON WEB SEARCH AND DATA MINING
(2021)
Proceedings Paper
Computer Science, Information Systems
Justin Wood, Wei Wang, Corey Arnold
Summary: This paper introduces a new interpretation of nonparametric Bayesian learning called the biased coin flip process, proving its equivalence to the Dirichlet process and demonstrating improved predictive performance.
DOCUMENT ANALYSIS AND RECOGNITION - ICDAR 2021, PT II
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Junheng Hao, Chuan Lei, Vasilis Efthymiou, Abdul Quamar, Fatma Ozcan, Yizhou Sun, Wei Wang
Summary: Medical ontologies and databases often have discrepancies that compromise interoperability, requiring data to ontology matching. Existing solutions focus on extracting information from ontologies for engineering, which can be labor-intensive. The proposed MEDTO framework utilizes three innovative techniques to achieve significant improvements in data to ontology matching.
KDD '21: PROCEEDINGS OF THE 27TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY & DATA MINING
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Zijie Huang, Yizhou Sun, Wei Wang
Summary: Many real-world systems are dynamic in nature, where coupled objects interact through graphs and exhibit complex behavior over time. The COVID-19 pandemic can be seen as a dynamic system with geographical locations as objects, influencing each other's infection rates. There is a need to explore how to accurately model and predict the complex dynamics of these systems.
KDD '21: PROCEEDINGS OF THE 27TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY & DATA MINING
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Cheng Zheng, Bo Zong, Wei Cheng, Dongjin Song, Jingchao Ni, Wenchao Yu, Haifeng Chen, Wei Wang
Summary: The proposed TSNet model jointly learns temporal and structural features for node classification from sparsified temporal graphs, effectively extracting local features and optimizing node representations to improve performance in node classification tasks.
MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES, ECML PKDD 2020, PT III
(2021)
Article
Biology
Xudong Zhu, Joseph M. Plasek, Chunlei Tang, Wasim Al-Assad, Zhikun Zhang, Yun Xiong, Liqin Wang, Sharmitha Yerneni, Carlos Ortega, Min-Jeoung Kang, Li Zhou, David W. Bates, Patricia C. Dykes
Summary: This study focuses on exploring and developing analysis tools for clinical notes, demonstrating how global embeddings, aligning at specific times, timeline reconstruction, and clustering can enhance representation learning and understanding of data connections in clinical documentation. The appropriate exploratory analysis tools not only improve data processing capabilities but also make data-driven medicine possible by providing keen insights into preprocessing clinical notes.
BMC RESEARCH NOTES
(2021)