MicroRNAs and complex diseases: from experimental results to computational models

Briefings in Bioinformatics, Год журнала: 2017, Номер 20(2), С. 515 - 539

Опубликована: Сен. 15, 2017

Plenty of microRNAs (miRNAs) were discovered at a rapid pace in plants, green algae, viruses and animals. As one the most important components cell, miRNAs play growing role various essential biological processes. For recent few decades, amounts experimental methods computational models have been designed implemented to identify novel miRNA-disease associations. In this review, functions miRNAs, miRNA-target interactions, associations some publicly available miRNA-related databases discussed detail. Specially, considering fact that an increasing number experimentally confirmed, we selected five human diseases crucial disease-related provided corresponding introductions. Identifying has become goal biomedical research, which will accelerate understanding disease pathogenesis molecular level tools design for diagnosis, treatment prevention. Computational means association identification, could select promising pairs validation significantly reduce time cost experiments. Here, reviewed 20 state-of-the-art predicting from different perspectives. Finally, summarized four factors difficulties potential framework constructing powerful predict including feasible research schemas, future directions further development models.

Язык: Английский

---

BioSeq-BLM: a platform for analyzing DNA, RNA and protein sequences based on biological language models

Nucleic Acids Research, Год журнала: 2021, Номер 49(22), С. e129 - e129

Опубликована: Сен. 9, 2021

In order to uncover the meanings of 'book life', 155 different biological language models (BLMs) for DNA, RNA and protein sequence analysis are discussed in this study, which able extract linguistic properties life'. We also extend BLMs into a system called BioSeq-BLM automatically representing analyzing data. Experimental results show that predictors generated by achieve comparable or even obviously better performance than exiting state-of-the-art published literatures, indicating will provide new approaches based on natural processing technologies, contribute development very important field. help readers use their own experiments, corresponding web server stand-alone package established released, can be freely accessed at http://bliulab.net/BioSeq-BLM/.

Язык: Английский

---

Circular RNAs and complex diseases: from experimental results to computational models

Briefings in Bioinformatics, Год журнала: 2021, Номер 22(6)

Опубликована: Июль 5, 2021

Abstract Circular RNAs (circRNAs) are a class of single-stranded, covalently closed RNA molecules with variety biological functions. Studies have shown that circRNAs involved in processes and play an important role the development various complex diseases, so identification circRNA-disease associations would contribute to diagnosis treatment diseases. In this review, we summarize discovery, classifications functions introduce four diseases associated circRNAs. Then, list some significant publicly accessible databases containing comprehensive annotation resources experimentally validated associations. Next, state-of-the-art computational models for predicting novel divide them into two categories, namely network algorithm-based machine learning-based models. Subsequently, several evaluation methods prediction performance these summarized. Finally, analyze advantages disadvantages different types provide suggestions promote association from perspective construction new accumulation circRNA-related data.

Язык: Английский

---

GMNN2CD: identification of circRNA–disease associations based on variational inference and graph Markov neural networks

Bioinformatics, Год журнала: 2022, Номер 38(8), С. 2246 - 2253

Опубликована: Фев. 9, 2022

Abstract Motivation With the analysis of characteristic and function circular RNAs (circRNAs), people have realized that they play a critical role in diseases. Exploring relationship between circRNAs diseases is far-reaching significance for searching etiopathogenesis treatment Nevertheless, it inefficient to learn new associations only through biotechnology. Results Consequently, we present computational method, GMNN2CD, which employs graph Markov neural network (GMNN) algorithm predict unknown circRNA–disease associations. First, used verified associations, calculate semantic similarity Gaussian interactive profile kernel (GIPs) disease GIPs circRNA then merge them form unified descriptor. After that, GMNN2CD uses fusion feature variational map autoencoder deep features label propagation propagate tags based on known Based inference, GMNN alternate training enhances ability obtain high-efficiency high-dimensional from low-dimensional representations. Finally, 5-fold cross-validation five benchmark datasets shows superior state-of-the-art methods. Furthermore, case studies shown can detect potential Availability implementation The source code data are available at https://github.com/nmt315320/GMNN2CD.git.

Язык: Английский

---

Exploring associations of non-coding RNAs in human diseases via three-matrix factorization with hypergraph-regular terms on center kernel alignment

Briefings in Bioinformatics, Год журнала: 2020, Номер 22(5)

Опубликована: Дек. 14, 2020

Relationship of accurate associations between non-coding RNAs and diseases could be great help in the treatment human biomedical research. However, traditional technology is only applied on one type RNA or a specific disease, experimental method time-consuming expensive. More computational tools have been proposed to detect new based known ncRNA disease information. Due ncRNAs (circRNAs, miRNAs lncRNAs) having close relationship with progression various diseases, it critical for developing effective predictors ncRNA-disease association prediction. In this paper, we propose three-matrix factorization hypergraph regularization terms (HGRTMF) central kernel alignment (CKA), identifying general associations. process constructing similarity matrix, types matrices are applicable circRNAs, lncRNAs. Our achieves excellent performance five datasets, involving three ncRNAs. test, obtain best area under curve scores $0.9832$, $0.9775$, $0.9023$, $0.8809$ $0.9185$ via 5-fold cross-validation $0.9836$, $0.9198$, $0.9459$ $0.9275$ leave-one-out datasets. Furthermore, our novel (CKA-HGRTMF) also able discover accurately. Availability: Codes data available: https://github.com/hzwh6910/ncRNA2Disease.git. Contact:[email protected].

Язык: Английский

---

A representation learning model based on variational inference and graph autoencoder for predicting lncRNA-disease associations

BMC Bioinformatics, Год журнала: 2021, Номер 22(1)

Опубликована: Март 21, 2021

Numerous studies have demonstrated that long non-coding RNAs are related to plenty of human diseases. Therefore, it is crucial predict potential lncRNA-disease associations for disease prognosis, diagnosis and therapy. Dozens machine learning deep algorithms been adopted this problem, yet still challenging learn efficient low-dimensional representations from high-dimensional features lncRNAs diseases unknown accurately.We proposed an end-to-end model, VGAELDA, which integrates variational inference graph autoencoders prediction. VGAELDA contains two kinds autoencoders. Variational (VGAE) infer respectively, while propagate labels via known associations. These trained alternately by adopting expectation maximization algorithm. The integration both the VGAE representation learning, alternate training inference, strengthens capability capture features, hence promotes robustness preciseness predicting Further analysis illuminates designed co-training framework lncRNA solves a geometric matrix completion problem capturing approach.Cross validations numerical experiments illustrate outperforms current state-of-the-art methods in association Case indicate capable detecting source code data available at https://github.com/zhanglabNKU/VGAELDA .

Язык: Английский

---

KGANCDA: predicting circRNA-disease associations based on knowledge graph attention network

Briefings in Bioinformatics, Год журнала: 2021, Номер 23(1)

Опубликована: Окт. 27, 2021

Abstract Increasing evidences have proved that circRNA plays a significant role in the development of many diseases. In addition, researches shown can be considered as potential biomarker for clinical diagnosis and treatment disease. Some computational methods been proposed to predict circRNA-disease associations. However, performance these is limited sparsity low-order interaction information. this paper, we propose new method (KGANCDA) associations based on knowledge graph attention network. The graphs are constructed by collecting multiple relationship data among circRNA, disease, miRNA lncRNA. Then, network designed obtain embeddings each entity distinguishing importance information from neighbors. Besides neighbor information, it also capture high-order multisource associations, which alleviates problem sparsity. Finally, multilayer perceptron applied affinity score experiment results show KGANCDA outperforms than other state-of-the-art 5-fold cross validation. Furthermore, case study demonstrates an effective tool

Язык: Английский

---

MGRCDA: Metagraph Recommendation Method for Predicting CircRNA–Disease Association

IEEE Transactions on Cybernetics, Год журнала: 2021, Номер 53(1), С. 67 - 75

Опубликована: Июль 8, 2021

Clinical evidence began to accumulate, suggesting that circRNAs can be novel therapeutic targets for various diseases and play a critical role in human health. However, limited by the complex mechanism of circRNA, it is difficult quickly large-scale explore relationship between disease circRNA wet-lab experiment. In this work, we design new computational model MGRCDA on account metagraph recommendation theory predict potential circRNA-disease associations. Specifically, first regard association prediction problem as system problem, series metagraphs according heterogeneous biological networks; then extract semantic information Gaussian interaction profile kernel (GIPK) similarity network attributes; finally, iterative search algorithm used calculate scores pair. On gold standard dataset circR2Disease, achieved accuracy 92.49% with an area under ROC curve 0.9298, which significantly higher than other state-of-the-art models. Furthermore, among top 30 disease-related recommended model, 25 have been verified latest published literature. The experimental results prove feasible efficient, recommend reliable candidates further experiment reduce scope

Язык: Английский

---

Predicting miRNA-Disease Association Based on Neural Inductive Matrix Completion with Graph Autoencoders and Self-Attention Mechanism

Biomolecules, Год журнала: 2022, Номер 12(1), С. 64 - 64

Опубликована: Янв. 2, 2022

Many studies have clarified that microRNAs (miRNAs) are associated with many human diseases. Therefore, it is essential to predict potential miRNA-disease associations for disease pathogenesis and treatment. Numerous machine learning deep approaches been adopted this problem. In paper, we propose a Neural Inductive Matrix completion-based method Graph Autoencoders (GAE) Self-Attention mechanism prediction (NIMGSA). Some of the previous works based on matrix completion ignore importance label propagation procedure inferring associations, while others cannot integrate effectively. Varying from studies, NIMGSA unifies inductive via neural network architecture, through collaborative training two graph autoencoders. This also an implementation self-attention prediction. end-to-end framework can strengthen robustness preciseness both propagation. Cross validations indicate outperforms current methods. Case demonstrate competent in detecting associations.

Язык: Английский

---

MPCLCDA: predicting circRNA–disease associations by using automatically selected meta-path and contrastive learning

Briefings in Bioinformatics, Год журнала: 2023, Номер 24(4)

Опубликована: Май 31, 2023

Circular RNA (circRNA) is closely associated with human diseases. Accordingly, identifying the associations between diseases and circRNA can help in disease prevention, diagnosis treatment. Traditional methods are time consuming laborious. Meanwhile, computational models effectively predict potential circRNA-disease (CDAs), but restricted by limited data, resulting data high dimension imbalance. In this study, we propose a model based on automatically selected meta-path contrastive learning, called MPCLCDA model. First, constructs new heterogeneous network similarity, similarity known association, via obtains low-dimensional fusion features of nodes graph convolutional networks. Then, learning used to optimize further, obtain node that make distinction positive negative samples more evident. Finally, scores predicted through multilayer perceptron. The proposed method compared advanced four datasets. average area under receiver operating characteristic curve, precision-recall curve F1 score 5-fold cross-validation reached 0.9752, 0.9831 0.9745, respectively. Simultaneously, case studies further prove predictive ability application value method.

Язык: Английский

---