Search

Analysis of miRNA-target mRNA interaction (MTI) is of crucial significance in discovering new target candidates for miRNAs. However, the biological experiments for identifying MTIs have a high false positive rate and are high-priced, time-consuming, and arduous. It is an urgent task to develop effective computational approaches to enhance the investigation of miRNA-target mRNA relationships. In this study, a novel method called MIPDH is developed for miRNA–mRNA interaction prediction by using DeepWalk on a heterogeneous network. More specifically, MIPDH extracts two kinds of features, in which a biological behavior feature is learned using a network embedding algorithm on a constructed heterogeneous network derived from 17 kinds of associations among drug, disease, and 6 kinds of biomolecules, and the attribute feature is learned using the k-mer method on sequences of miRNAs and target mRNAs. Then, a random forest classifier is trained on the features combined with the biological behavior feature and attribute feature. When implementing a 5-fold cross-validation experiment, MIPDH achieved an average accuracy, sensitivity, specificity and AUC of 75.85, 74.37, 77.33%, and 0.8044, respectively. To further evaluate the performance of MIPDH, other classifiers and feature descriptors are conducted for comparisons. MIPDH can achieve a better performance. Additionally, case studies on hsa-miR-106b-5p, hsa-let-7d-5p, and hsa-let-7e-5p are also implemented. As a result, 14, 9, and 9 out of the top 15 targets that interacted with these miRNAs were verified using the experimental literature or other databases. All these prediction results indicate that MIPDH is an effective method for predicting miRNA-target mRNA interactions.

Effectively representing Medical Subject Headings (MeSH) headings (terms) such as disease and drug as discriminative vectors could greatly improve the performance of downstream computational prediction models. However, these terms are often abstract and difficult to quantify. In this paper, we converted the MeSH tree structure into a relationship network and applied several graph embedding algorithms on it to represent these terms. Specifically, the relationship network consisting of nodes (MeSH headings) and edges (relationships), which can be constructed by the tree num. Then, five graph embedding algorithms including DeepWalk, LINE, SDNE, LAP and HOPE were implemented on the relationship network to represent MeSH headings as vectors. In order to evaluate the performance of the proposed methods, we carried out the node classification and relationship prediction tasks. The results show that the MeSH headings characterized by graph embedding algorithms can not only be treated as an independent carrier for representation, but also can be utilized as additional information to enhance the representation ability of vectors. Thus, it can serve as an input and continue to play a significant role in any computational models related to disease, drug, microbe, etc. Besides, our method holds great hope to inspire relevant researchers to study the representation of terms in this network perspective.

Graphical abstract Complex biological network based on known molecular association structure., Benefiting from advances in high-throughput experimental techniques, important regulatory roles of miRNAs, lncRNAs, and proteins, as well as biological property information, are gradually being complemented. As the key data support to promote biomedical research, domain knowledge such as intermolecular relationships that are increasingly revealed by molecular genome-wide analysis is often used to guide the discovery of potential associations. However, the method of performing network representation learning from the perspective of the global biological network is scarce. These methods cover a very limited type of molecular associations and are therefore not suitable for more comprehensive analysis of molecular network representation information. In this study, we propose a computational model based on the Biological network for predicting potential associations between miRNAs and diseases called iMDA-BN. The iMDA-BN has three significant advantages: I) It uses a new method to describe disease and miRNA characteristics which analyzes node representation information for disease and miRNA from the perspective of biological networks. II) It can predict unproven associations even if miRNAs and diseases do not appear in the biological network. III) Accurate description of miRNA characteristics from biological properties based on high-throughput sequence information. The iMDA-BN predictor achieves an AUC of 0.9145 and an accuracy of 84.49% on the miRNA-disease association baseline dataset, and it can also achieve an AUC of 0.8765 and an accuracy of 80.96% when predicting unknown diseases and miRNAs in the biological network. Compared to existing miRNA-disease association prediction methods, iMDA-BN has higher accuracy and the advantage of predicting unknown associations. In addition, 45, 49, and 49 of the top 50 miRNA-disease associations with the highest predicted scores were confirmed in the case studies, respectively.

Detecting whether a pair of biomolecules associate is of great significance in the study of molecular biology. Hence, computational methods are urgently needed as guidance for practice. However, most of the previous prediction models influenced by reductionism focused on isolated research objects, which have their own inherent defects. Inspired by holism, a machine-learning-based framework called MAN-node2vec is proposed to predict multi-type relationships in the molecular associations network (MAN). Specifically, we constructed a large-scale MAN composed of 1,023 miRNAs, 1,649 proteins, 769 long non-coding RNAs (lncRNAs), 1,025 drugs, and 2,062 diseases. Then, each biomolecule in MAN can be represented as a vector by its attribute learned by k-mer, etc. and its behavior learned by node2vec. Finally, the random forest classifier is applied to carry out the relationship prediction task. The proposed model achieved a reliable performance with 0.9677 areas under the curve (AUCs) and 0.9562 areas under the precision curve (AUPRs) under 5-fold cross-validation. Also, additional experiments proved that the proposed global model shows more competitive performance than the traditional local method. All of these provided a systematic insight for understanding the synergistic interactions between various molecules and diseases. It is anticipated that this work can bring beneficial inspiration and advance to related systems biology and biomedical research.

Identification of drug-target interactions (DTIs) is a significant step in the drug discovery or repositioning process. Compared with the time-consuming and labor-intensive in vivo experimental methods, the computational models can provide high-quality DTI candidates in an instant. In this study, we propose a novel method called LGDTI to predict DTIs based on large-scale graph representation learning. LGDTI can capture the local and global structural information of the graph. Specifically, the first-order neighbor information of nodes can be aggregated by the graph convolutional network (GCN), on the other hand, the high-order neighbor information of nodes can be learned by the graph embedding method called DeepWalk. Finally, the two kinds of feature are fed into the random forest classifier to train and predict potential DTIs. The results show that our method obtained area under the receiver operating characteristic curve (AUROC) of 0.9455 and area under the precision-recall curve (AUPR) of 0.9491 under 5-fold cross-validation. Moreover, we compare the presented method with some existing state-of-the-art methods. These results imply that LGDTI can efficiently and robustly capture undiscovered DTIs. Moreover, the proposed model is expected to bring new inspiration and provide novel perspectives to relevant researchers.

Background The prediction of potential drug-target interactions (DTIs) not only provides a better comprehension of biological processes but also is critical for identifying new drugs. However, due to the disadvantages of expensive and high time-consuming traditional experiments, only a small section of interactions between drugs and targets in the database were verified experimentally. Therefore, it is meaningful and important to develop new computational methods with good performance for DTIs prediction. At present, many existing computational methods only utilize the single type of interactions between drugs and proteins without paying attention to the associations and influences with other types of molecules. Methods In this work, we developed a novel network embedding-based heterogeneous information integration model to predict potential drug-target interactions. Firstly, a heterogeneous multi-molecuar information network is built by combining the known associations among protein, drug, lncRNA, disease, and miRNA. Secondly, the Large-scale Information Network Embedding (LINE) model is used to learn behavior information (associations with other nodes) of drugs and proteins in the network. Hence, the known drug-protein interaction pairs can be represented as a combination of attribute information (e.g. protein sequences information and drug molecular fingerprints) and behavior information of themselves. Thirdly, the Random Forest classifier is used for training and prediction. Results In the results, under the five-fold cross validation, our method obtained 85.83% prediction accuracy with 80.47% sensitivity at the AUC of 92.33%. Moreover, in the case studies of three common drugs, the top 10 candidate targets have 8 (Caffeine), 7 (Clozapine) and 6 (Pioglitazone) are respectively verified to be associated with corresponding drugs. Conclusions In short, these results indicate that our method can be a powerful tool for predicting potential drug-target interactions and finding unknown targets for certain drugs or unknown drugs for certain targets.

Background: The prediction of potential drug-protein target interactions (DTIs) not only provides a better comprehension of biological processes but also is critical for identifying new drugs. However, due to the disadvantages of expensive and high time-consuming traditional experiments, only a small section of interactions between drugs and targets in the database were verified experimentally. Therefore, it is meaningful and important to develop new computational methods with good performance for DTIs prediction. At present, many existing computational methods only utilize the single type of interactions between drugs and proteins without paying attention to the associations and influences with other types of molecules. Methods: In this work, we developed a novel network embedding-based heterogeneous information integration model to predict potential drug-target interactions. Firstly, a heterogeneous information network is built by combining the known associations among protein, drug, lncRNA, disease, and miRNA. Secondly, the Large-scale Information Network Embedding (LINE) model is used to learn behavior information (associations with other nodes) of drugs and proteins in the network. Hence, the known drug-protein interaction pairs can be represented as a combination of attribute information (e.g. protein sequences information and drug molecular fingerprints) and behavior information of themselves. Thirdly, the Random Forest classifier is used for training and prediction. Results: In the results, under the 5-fold cross validation, our method obtained 85.83% prediction accuracy with 80.47% sensitivity at the AUC of 92.33%. Moreover, in the case studies of three common drugs, the top 10 candidate targets have 8 (Caffeine), 7 (Clozapine) and 6 (Pioglitazone) are respectively verified to be associated with corresponding drugs. Conclusions: In short, these results indicate that our method can be a powerful tool for predicting potential drug-protein interactions and finding unknown targets for certain drugs or unknown drugs for certain targets.

Abundant life activities are maintained by various biomolecule relationships in human cells. However, many previous computational models only focus on isolated objects, without considering that cell is a complete entity with ample functions. Inspired by holism, we constructed a Molecular Associations Network (MAN) including 9 kinds of relationships among 5 types of biomolecules, and a prediction model called MAN-GF. More specifically, biomolecules can be represented as vectors by the algorithm called biomarker2vec which combines 2 kinds of information involved the attribute learned by k-mer, etc and the behavior learned by Graph Factorization (GF). Then, Random Forest classifier is applied for training, validation and test. MAN-GF obtained a substantial performance with AUC of 0.9647 and AUPR of 0.9521 under 5-fold Cross-validation. The results imply that MAN-GF with an overall perspective can act as ancillary for practice. Besides, it holds great hope to provide a new insight to elucidate the regulatory mechanisms., Guo et al. construct a large scale Molecular Associations Network (MAN) including 9 kinds of associations among 5 types of biomolecules, namely protein, miRNA, lncRNA, disease and drug. They further propose a computational model, MAN-GF, that can predict links between these biomolecules and displays a substantial performance under 5-fold cross validation.

A key aim of post-genomic biomedical research is to systematically understand molecules and their interactions in human cells. Multiple biomolecules coordinate to sustain life activities, and interactions between various biomolecules are interconnected. However, existing studies usually only focusing on associations between two or very limited types of molecules. In this study, we propose a network representation learning based computational framework MAN-SDNE to predict any intermolecular associations. More specifically, we constructed a large-scale molecular association network of multiple biomolecules in human by integrating associations among long non-coding RNA, microRNA, protein, drug, and disease, containing 6,528 molecular nodes, 9 kind of,105,546 associations. And then, the feature of each node is represented by its network proximity and attribute features. Furthermore, these features are used to train Random Forest classifier to predict intermolecular associations. MAN-SDNE achieves a remarkable performance with an AUC of 0.9552 and an AUPR of 0.9338 under five-fold cross-validation. To indicate the ability to predict specific types of interactions, a case study for predicting lncRNA-protein interactions using MAN-SDNE is also executed. Experimental results demonstrate this work offers a systematic insight for understanding the synergistic associations between molecules and complex diseases and provides a network-based computational tool to systematically explore intermolecular interactions.

Background The interactions between non-coding RNAs (ncRNA) and proteins play an essential role in many biological processes. Several high-throughput experimental methods have been applied to detect ncRNA-protein interactions. However, these methods are time-consuming and expensive. Accurate and efficient computational methods can assist and accelerate the study of ncRNA-protein interactions. Results In this work, we develop a stacking ensemble computational framework, RPI-SE, for effectively predicting ncRNA-protein interactions. More specifically, to fully exploit protein and RNA sequence feature, Position Weight Matrix combined with Legendre Moments is applied to obtain protein evolutionary information. Meanwhile, k-mer sparse matrix is employed to extract efficient feature of ncRNA sequences. Finally, an ensemble learning framework integrated different types of base classifier is developed to predict ncRNA-protein interactions using these discriminative features. The accuracy and robustness of RPI-SE was evaluated on three benchmark data sets under five-fold cross-validation and compared with other state-of-the-art methods. Conclusions The results demonstrate that RPI-SE is competent for ncRNA-protein interactions prediction task with high accuracy and robustness. It’s anticipated that this work can provide a computational prediction tool to advance ncRNA-protein interactions related biomedical research.

Recent studies have shown that lncRNAs play a critical role in numerous complex human diseases. Thus, identification of lncRNA and diseases associations can help us to understand disease pathogenesis at the molecular level and develop disease diagnostic biomarkers. In this paper, a novel computational method LDAMAN is proposed to predict potential lncRNA-disease interactions from heterogeneous information network with SDNE embedding model. Specifically, known associations among lncRNA, disease, microRNA, circular RNA, mRNA, protein, drug and microbe are integrated to construct a molecular association network and a network embedding model SDNE is employed to extract network behavior features of lncRNA and disease nodes. Finally, the XGBoost classifier is used for predicting potential lncRNA-disease associations. In the experiment, the proposed method obtained stable AUC of 92.58% using 5-fold cross validation. In summary, the experimental results demonstrate our method provides a systematic landscape and computational prediction tool for lncRNA-disease association prediction.

Accurate identification of drug-target interactions (DTIs) is essential for drug development. It not only helps the researchers to understand the mechanism of drug action, but also contributes to the innovative drug discovery and repositioning. However, due to the limitation the high cost and long time, the traditional experimental methods are difficult to be widely applied for DTIs prediction. In this study, we propose an in silico method for predicting drug-target interactions by Node2vec node embedding in molecular associations network (MAN). Specifically, the MAN is constructed by integrating the associations among drug, protein, disease, lncRNA and miRNA. Then, the node2vec embedding method is employed to obtain a behavior feature vector of each node in the network. The traditional attribute feature vector comes from the drug molecular fingerprint and protein sequences. Finally, a random forest (RF) classifier is performed on these features to predict potential drug-target pairs. The experimental results show that the behavior feature could obtain 87.37% accuracy, which is obviously better than the traditional attribute feature. This work is not only more robust and reliable for predicting DTIs, but also provides an alternative way for other biomolecules associations prediction.

Identification of drug-disease association is crucial for drug development and reposition. However, discovering drugs which are associated with diseases from in vitro testing is costly and time-consuming. Accumulating evidence showed that computational approaches can complement biological and clinical experiments for this identification task. In this work, we propose a novel computational method Node2Bio for predicting drug-disease associations using a highly efficient biomolecular network representation model. Specifically, we first construct a large-scale biomolecular association network (BAN) by integrating the associations among drugs, diseases, proteins, miRNAs and lncRNAs. Then, the network embedding model node2vec is used to extract network behavior features of drug and disease nodes. Finally, the feature vectors are taken as inputs for the XGboost classifier to predict potential drug-disease associations. To evaluate the prediction performance of the proposed method, five-fold cross-validation tests are performed on a widely used SCMFDD-S dataset. The experimental results demonstrate that our method achieves competitive performance with a high AUC value of 0.8569, which suggests that our method is a useful tool for identification of drug-disease associations.

The accumulated evidences indicate that drugs not only interact with proteins, but also regulate a wide variety of biomarkers such as miRNAs. Hence, uncovering potential drug-miRNA associations plays significant roles in disease prevention, diagnosis and treatment as well as drug development. In this paper, we discuss how this problem is formulated as a link prediction task in a bipartite graph and construct a computational model to infer unknown drug-miRNA associations. Specifically, the drug SMILES (Simplified molecular input line entry specification) or miRNA sequences can be regarded as a kind of biology language described by distributed representation. The experiment verified associations are treated as positive samples and the same number unlabeled associations are randomly selected as negative samples. Finally, Random Forest classifier is applied to perform the prediction task. In the experiment, the proposed method achieves AUROC of 91.16 and AUPR of 89.21 under 5-fold cross-validation. It demonstrates the great potential of seamless integration of deep learning and biological big data. We hope that this research with great expectations can be used as a practical guidance tool to bring useful inspiration to relevant researchers.

Machine learning methods are increasingly being applied to model and predict biomolecular interactions, while efficient feature representation plays a vital role. To this end, a unified biological sequence deep representation learning framework BioSeq2vec is proposed to extract discriminative features of any type of biological sequence. For arbitrary-length sequence input, the BioSeq2vec produces fixed-length efficient feature representation, which can be applied to various learning models. The performance of BioSeq2vec is evaluated on lncRNA-protein interaction prediction tasks. Experimental results reveal the superior performance of BioSeq2vec in biological sequence feature representation and broad prospects in various genome informatics and computational biology studies.

DNA-binding proteins (DBPs) play vital roles in the regulation of biological systems. Although there are already many deep learning methods for predicting the sequence specificities of DBPs, they face two challenges as follows. Classic deep learning methods for DBPs prediction usually fail to capture the dependencies between genomic sequences since their commonly used one-hot codes are mutually orthogonal. Besides, these methods usually perform poorly when samples are inadequate. To address these two challenges, we developed a novel language model for mining DBPs using human genomic data and ChIP-seq datasets with decaying learning rates, named DNA Fine-tuned Language Model (DFLM). It can capture the dependencies between genome sequences based on the context of human genomic data and then fine-tune the features of DBPs tasks using different ChIP-seq datasets. First, we compared DFLM with the existing widely used methods on 69 datasets and we achieved excellent performance. Moreover, we conducted comparative experiments on complex DBPs and small datasets. The results show that DFLM still achieved a significant improvement. Finally, we found that one-hot encoding completely cut off the dependencies of DNA sequences themselves, while DFLM using language models can well represent the dependency of DNA sequences.

The explosive growth of genomic, chemical and pathological data provides new opportunities and challenges to re-recognize life activities within human cells. However, there exist few computational models that aggregate various biomarkers to comprehensively reveal the physical and functional landscape of the biology system. Here, we construct a graph called Molecular Association Network (MAN) and a representation method called Biomarker2vec. Specifically, MAN is a heterogeneous attribute network consists of 18 kinds of edges (relationships) among 8 kinds of nodes (biomarkers). Biomarker2vec is an algorithm that represents the nodes as vectors by integrating biomarker attribute and behavior. After the biomarkers are described as vectors, random forest classifier is applied to carry out the prediction task. Our approach achieved promising performance on 18 relationships, with AUC of 0.9608 and AUPR of 0.9572. We also empirically explored the contribution of attribute and behavior feature of biomarkers to the results. In addition, a drug-disease association prediction case study was performed to validate our method’s ability on a specific object. These results strongly prove that MAN is a network with rich topological and biological information and Biomarker2vec can indeed adequately characterize biomarkers. Generally, our method can achieve simultaneous prediction of both single-type and multi-type relationships, which bring beneficial inspiration to relevant scholars and expand the medical research paradigm.

MotivationEffectively representing the MeSH headings (terms) such as disease and drug as discriminative vectors could greatly improve the performance of downstream computational prediction models. However, these terms are often abstract and difficult to quantify.ResultsIn this paper, we converted the MeSH tree structure into a relationship network and applied several graph embedding algorithms on it to represent these terms. Specifically, the relationship network consisting of nodes (MeSH headings) and edges (relationships) which can be constructed by the rule of tree num. Then, five graph embedding algorithms including DeepWalk (DW), LINE, SDNE, LAP and HOPE were implemented on the relationship network to represent MeSH headings as vectors. In order to evaluate the performance of the proposed method, we carried out the node classification and relationship prediction tasks. The experimental results show that the MeSH headings characterized by graph embedding algorithms can not only be treated as an independent carrier for representation, but also can be utilized as additional information to enhance the distinguishable ability of vectors. Thus, it can act as input and continue to play a significant role in any disease-, drug-, microbe- and etc.-related computational models. Besides, our method holds great hope to inspire relevant researchers to study the representation of terms in this network perspective.Contactzhuhongyou@ms.xjb.ac.cn

A key aim of post-genomic biomedical research is to systematically integrate and model all molecules and their interactions in living cells. Existing research usually only focusing on the associations between individual or very limited type of molecules. But the interactions between molecules shouldn’t be isolated but interconnected and influenced. In this study, we revealed, constructed and analyzed a large-scale molecular association network of multiple biomolecules inhumancells by modeling all associations among lncRNA, miRNA, protein, circRNA, microbe, drug, and disease, in which various associations are interconnected and any type of associations can be predicted. More specifically, we defined the molecular associations network and constructed a molecular associations dataset containing 105546 associations. Then, each node is represented by its attribute feature and network embedding learned by Structural Deep Network Embedding. Moreover, Random Forest is trained to predict any kind of associations. And we compared the features and classifiers under five-fold cross-validation. Our method achieves a remarkable performance on entire molecular associations network with an AUC of 0.9552 and an AUPR of 0.9338. To further evaluate the performance of our method, a case study for predicting lncRNA-protein interactions was executed. The experimental results demonstrate that the systematic insight for understanding the synergistic interactions between various molecules and complex diseases. It is anticipated that this work can bring beneficial inspiration and advance related systems biology and biomedical research.Author SummaryThe interactions between the various biomolecules in the cells should not be isolated, but interconnected and influenced. There have been many valuable studies on the interactions between two individual molecules. Based on a systematic and holistic perspective, we revealed and constructed a large-scale molecular associations network by combining various associations in human living cells, including miRNA-lncRNA association, miRNA-disease association, miRNA-protein interaction, lncRNA-disease association, protein-protein interaction, protein-disease association, drug-protein interaction, drug-disease interaction, and lncRNA-protein interaction. To model and analyze this molecular associations network, we employed the network representation learning model to learn how to represent the node. Each node in the network can be represented by network embedding and its own attribute information. Any node can be classified. And any type of the associations in this network can be predicted, which can be considered as link prediction task. Our work provides a new systematic view and conceptual framework to understand complex diseases and life activities. It is anticipated that our study can advance related biological macromolecules, systems biology and biomedical research, and bring some meaningful inspiration.