Long non-coding RNAs and complex diseases: from experimental results to computational models

Sparse Representation-Based Patient-Specific Diagnosis and Treatment for Esophageal Squamous Cell Carcinoma

Precision medicine and personalized treatment have attracted attention in recent years. However, most genetic medicines mainly target one genetic site, while complex diseases like esophageal squamous cell carcinoma (ESCC) usually present heterogeneity that involves variations of many genetic markers. Here, we seek an approach to leverage genetic data and ESCC knowledge data to forward personalized diagnosis and treatment for ESCC. First, 851 ESCC-related gene markers and their druggability were studied through a comprehensive literature analysis. Then, a sparse representation-based variable selection (SRVS) was employed for patient-specific genetic marker selection using gene expression datasets. Results showed that the SRVS method could identify a unique gene vector for each patient group, leading to significantly higher classification accuracies compared to randomly selected genes (100, 97.17, 100, 100%; permutation p values: 0.0032, 0.0008, 0.0004, and 0.0008). The SRVS also outperformed an ANOVA-based gene selection method in terms of the classification ratio. The patient-specific gene markers are targets of ESCC effective drugs, providing specific guidance for medicine selection. Our results suggest the effectiveness of integrating previous database utilizing SRVS in assisting personalized medicine selection and treatment for ESCC.

Semi-Supervised Maximum Discriminative Local Margin for Gene Selection

In the present study, we introduce a novel semi-supervised method called the semi-supervised maximum discriminative local margin (semiMM) for gene selection in expression data. The semiMM is a "filter" approach that exploits local structure, variance, and mutual information. We first constructed a local nearest neighbour graph and divided this information into within-class and between-class local nearest neighbour graphs by weighing the edge between the two data points. The semiMM aims to discover the most discriminative features for classification via maximizing the local margin between the within-class and between-class data, the variance of all data, and the mutual information of features with class labels. Experiments on five publicly available gene expression datasets revealed the effectiveness of the proposed method compared to three state-of-the-art feature selection algorithms.

Analysis of long non-coding RNA (lncRNA) expression in hepatitis B patients

Long non-coding RNAs (lncRNAs) have been implicated in numerous biological processes, including epigenetic regulation, cell-cycle control, and transcriptional/translational regulation of gene expression. Differential expression of lncRNAs and disruption of the regulatory processes are recognized as critical steps in cancer development. The role of lncRNAs in hepatitis B virus (HBV) infection is not well understood. Here we analyzed the expression of 135 lncRNAs in plasma samples of 82 HBV patients (classified as chronic patients, inactive carriers, or resolved patients) at diagnosis and at 12 months of treatment in relation to control group (81 healthy volunteers). We also investigated the effect of small interfering RNA (siRNA)-mediated silencing of lincRNA-SFMBT2 on HBV-positive human liver cancer cell line. lncRNA expression was analyzed by quantitative reverse transcription-polymerase chain reaction (qRT-PCR). Chemically synthesized siRNAs were transfected into the cell lines using Lipofectamine 2000 Reagent (Thermo Fisher Scientific). HBV DNA and HBsAg and HBeAg were detected in transfected cultures by real-time PCR and ELISA, respectively, using commercial kits. We observed changes in lncRNA expression in all three HBV groups, compared to control group. Most notably, the expression of anti-NOS2A, lincRNA-SFMBT2, and Zfhx2as was significantly increased and expression of Y5 lncRNA was decreased in chronic HBV patients. A decreased Y5 expression and increased lincRNA-SFMBT2 expression were observed in inactive HBsAg carriers. The expression of HOTTIP, MEG9, and PCAT-32 was increased in resolved HBV patients, and no significant change in the expression of Y5 was observed, compared to control group. siRNA-mediated inhibition of lincRNA-SFMBT2 decreased the level of HBV DNA in human liver cancer cells. Further research is needed to confirm the prognostic as well as therapeutic role of these lncRNAs in HBV patients.

Model-based and Model-free Machine Learning Techniques for Diagnostic Prediction and Classification of Clinical Outcomes in Parkinson's Disease

In this study, we apply a multidisciplinary approach to investigate falls in PD patients using clinical, demographic and neuroimaging data from two independent initiatives (University of Michigan and Tel Aviv Sourasky Medical Center). Using machine learning techniques, we construct predictive models to discriminate fallers and non-fallers. Through controlled feature selection, we identified the most salient predictors of patient falls including gait speed, Hoehn and Yahr stage, postural instability and gait difficulty-related measurements. The model-based and model-free analytical methods we employed included logistic regression, random forests, support vector machines, and XGboost. The reliability of the forecasts was assessed by internal statistical (5-fold) cross validation as well as by external out-of-bag validation. Four specific challenges were addressed in the study: Challenge 1, develop a protocol for harmonizing and aggregating complex, multisource, and multi-site Parkinson's disease data; Challenge 2, identify salient predictive features associated with specific clinical traits, e.g., patient falls; Challenge 3, forecast patient falls and evaluate the classification performance; and Challenge 4, predict tremor dominance (TD) vs. posture instability and gait difficulty (PIGD). Our findings suggest that, compared to other approaches, model-free machine learning based techniques provide a more reliable clinical outcome forecasting of falls in Parkinson's patients, for example, with a classification accuracy of about 70-80%.

A Bipartite Network and Resource Transfer-Based Approach to Infer lncRNA-Environmental Factor Associations

Phenotypes and diseases are often determined by the complex interactions between genetic factors and environmental factors (EFs). However, compared with protein-coding genes and microRNAs, there is a paucity of computational methods for understanding the associations between long non-coding RNAs (lncRNAs) and EFs. In this study, we focused on the associations between lncRNA and EFs. By using the common miRNA partners of any pair of lncRNA and EF, based on the competing endogenous RNA (ceRNA) hypothesis and the technique of resources transfer within the experimentally-supported lncRNA-miRNA and miRNA-EF association bipartite networks, we propose an algorithm for predicting new lncRNA-EF associations. Results show that, compared with another recently-proposed method, our approach is capable of predicting more credible lncRNA-EF associations. These results support the validity of our approach to predict biologically significant associations, which could lead to a better understanding of the molecular processes.

LPI-ETSLP: lncRNA-protein interaction prediction using eigenvalue transformation-based semi-supervised link prediction

RNA-protein interactions are essential for understanding many important cellular processes. In particular, lncRNA-protein interactions play important roles in post-transcriptional gene regulation, such as splicing, translation, signaling and even the progression of complex diseases. However, the experimental validation of lncRNA-protein interactions remains time-consuming and expensive, and only a few theoretical approaches are available for predicting potential lncRNA-protein associations. Here, we presented eigenvalue transformation-based semi-supervised link prediction (LPI-ETSLP) to uncover the relationship between lncRNAs and proteins. Moreover, it is semi-supervised and does not need negative samples. Based on 5-fold cross validation, an AUC of 0.8876 and an AUPR of 0.6438 have demonstrated its reliable performance compared with three other computational models. Furthermore, the case study demonstrated that many lncRNA-protein interactions predicted by our method can be successfully confirmed by experiments. It is indicated that LPI-ETSLP would be a useful bioinformatics resource for biomedical research studies.

Heterogeneity Analysis and Diagnosis of Complex Diseases Based on Deep Learning Method

Understanding genetic mechanism of complex diseases is a serious challenge. Existing methods often neglect the heterogeneity phenomenon of complex diseases, resulting in lack of power or low reproducibility. Addressing heterogeneity when detecting epistatic single nucleotide polymorphisms (SNPs) can enhance the power of association studies and improve prediction performance of complex diseases diagnosis. In this study, we propose a three-stage framework including epistasis detection, clustering and prediction to address both epistasis and heterogeneity of complex diseases based on deep learning method. The epistasis detection stage applies a multi-objective optimization method to find several candidate sets of epistatic SNPs which contribute to different subtypes of complex diseases. Then, a K-means clustering algorithm is used to define subtypes of the case group. Finally, a deep learning model has been trained for disease prediction based on graphics processing unit (GPU). Experimental results on pure and heterogeneous datasets show that our method has potential practicality and can serve as a possible alternative to other methods. Therefore, when epistasis and heterogeneity exist at the same time, our method is especially suitable for diagnosis of complex diseases.

An Automatic Diagnosis Method of Facial Acne Vulgaris Based on Convolutional Neural Network

In this paper, we present a new automatic diagnosis method for facial acne vulgaris which is based on convolutional neural networks (CNNs). To overcome the shortcomings of previous methods which were the inability to classify enough types of acne vulgaris. The core of our method is to extract features of images based on CNNs and achieve classification by classifier. A binary-classifier of skin-and-non-skin is used to detect skin area and a seven-classifier is used to achieve the classification task of facial acne vulgaris and healthy skin. In the experiments, we compare the effectiveness of our CNN and the VGG16 neural network which is pre-trained on the ImageNet data set. We use a ROC curve to evaluate the performance of binary-classifier and use a normalized confusion matrix to evaluate the performance of seven-classifier. The results of our experiments show that the pre-trained VGG16 neural network is effective in extracting features from facial acne vulgaris images. And the features are very useful for the follow-up classifiers. Finally, we try applying the classifiers both based on the pre-trained VGG16 neural network to assist doctors in facial acne vulgaris diagnosis.

A Deep Learning Framework for Robust and Accurate Prediction of ncRNA-Protein Interactions Using Evolutionary Information

The interactions between non-coding RNAs (ncRNAs) and proteins play an important role in many biological processes, and their biological functions are primarily achieved by binding with a variety of proteins. High-throughput biological techniques are used to identify protein molecules bound with specific ncRNA, but they are usually expensive and time consuming. Deep learning provides a powerful solution to computationally predict RNA-protein interactions. In this work, we propose the RPI-SAN model by using the deep-learning stacked auto-encoder network to mine the hidden high-level features from RNA and protein sequences and feed them into a random forest (RF) model to predict ncRNA binding proteins. Stacked assembling is further used to improve the accuracy of the proposed method. Four benchmark datasets, including RPI2241, RPI488, RPI1807, and NPInter v2.0, were employed for the unbiased evaluation of five established prediction tools: RPI-Pred, IPMiner, RPISeq-RF, lncPro, and RPI-SAN. The experimental results show that our RPI-SAN model achieves much better performance than other methods, with accuracies of 90.77%, 89.7%, 96.1%, and 99.33%, respectively. It is anticipated that RPI-SAN can be used as an effective computational tool for future biomedical researches and can accurately predict the potential ncRNA-protein interacted pairs, which provides reliable guidance for biological research.

Prediction and characterization of human ageing-related proteins by using machine learning

Ageing has a huge impact on human health and economy, but its molecular basis - regulation and mechanism - is still poorly understood. By today, more than three hundred genes (almost all of them function as protein-coding genes) have been related to human ageing. Although individual ageing-related genes or some small subsets of these genes have been intensively studied, their analysis as a whole has been highly limited. To fill this gap, for each human protein we extracted 21000 protein features from various databases, and using these data as an input to state-of-the-art machine learning methods, we classified human proteins as ageing-related or non-ageing-related. We found a simple classification model based on only 36 protein features, such as the "number of ageing-related interaction partners", "response to oxidative stress", "damaged DNA binding", "rhythmic process" and "extracellular region". Predicted values of the model quantify the relevance of a given protein in the regulation or mechanisms of the human ageing process. Furthermore, we identified new candidate proteins having strong computational evidence of their important role in ageing. Some of them, like Cytochrome b-245 light chain (CY24A) and Endoribonuclease ZC3H12A (ZC12A) have no previous ageing-associated annotations.

HGIMDA: Heterogeneous graph inference for miRNA-disease association prediction

Recently, microRNAs (miRNAs) have drawn more and more attentions because accumulating experimental studies have indicated miRNA could play critical roles in multiple biological processes as well as the development and progression of human complex diseases. Using the huge number of known heterogeneous biological datasets to predict potential associations between miRNAs and diseases is an important topic in the field of biology, medicine, and bioinformatics. In this study, considering the limitations in the previous computational methods, we developed the computational model of Heterogeneous Graph Inference for MiRNA-Disease Association prediction (HGIMDA) to uncover potential miRNA-disease associations by integrating miRNA functional similarity, disease semantic similarity, Gaussian interaction profile kernel similarity, and experimentally verified miRNA-disease associations into a heterogeneous graph. HGIMDA obtained AUCs of 0.8781 and 0.8077 based on global and local leave-one-out cross validation, respectively. Furthermore, HGIMDA was applied to three important human cancers for performance evaluation. As a result, 90% (Colon Neoplasms), 88% (Esophageal Neoplasms) and 88% (Kidney Neoplasms) of top 50 predicted miRNAs are confirmed by recent experiment reports. Furthermore, HGIMDA could be effectively applied to new diseases and new miRNAs without any known associations, which overcome the important limitations of many previous computational models.

Which statistical significance test best detects oncomiRNAs in cancer tissues? An exploratory analysis

MicroRNAs(miRNAs) often exert their oncogenic and tumor suppressor functions by suppressing protein-coding genes expressions in cancers and thus have a strong association with cancers' generation, development and metastasis. Through comprehensively understanding differentially expressed miRNAs (oncomiRNA) in tumor tissues, we can elucidate the underlying molecular mechanisms in tumorigenesis and develop novel strategies for cancer diagnosis and treatment. The differential expression of miRNAs can now be analyzed through numerous statistical significance tests based on different principles, which are also available in various R packages. However, the results can be notably different. In this study, we compared miRNAs obtained from 6 common significance tests/R packages (t-test, Limma, DESeq, edgeR, LRT and MARS) with the miRNAs archived in two databases; HMDD 2.0 database, which collects experimentally validated differentially expressed miRNAs, and Infer microRNA-disease association database, which contains the potential disease-associated miRNAs by network forecasting. Finally, we sought the MARS method in DEGseq package more effectively searched out differentially expressed miRNAs than other common methods.

H19 knockdown suppresses proliferation and induces apoptosis by regulating miR-148b/WNT/β-catenin in ox-LDL -stimulated vascular smooth muscle cells

Background

Long non-coding RNAs (lncRNAs) have been identified as critical regulators in the development of atherosclerosis (AS). Here, we focused on discussing roles and molecular mechanisms of lncRNA H19 in vascular smooth muscle cells (VSMCs) progression.

Methods

RT-qPCR assay was used to detect the expression patterns of H19 and miR-148b in clinical samples and cells. Cell proliferative ability was evaluated by CCK-8 and colony formation assays. Cell apoptotic capacity was assessed by apoptotic cell percentage and the caspase-3 activity. Bioinformatics analysis, luciferase and RNA immunoprecipitation (RIP) assays were employed to demonstrate cell percentage and the relationship among H19, miR-148b and wnt family member 1 (WNT1). Western blot assay was performed to determine expressions of proliferating cell nuclear antigen (PCNA), ki-67, Bax, Bcl-2, WNT1, β-catenin, C-myc and E-cadherin.

Results

The level of H19 was increased and miR-148b expression was decreased in human AS patient serums and oxidized low-density lipoprotein (ox-LDL)-stimulated human aorta vascular smooth muscle cells (HA-VSMCs). H19 knockdown suppressed proliferation and promoted apoptosis in HA-VSMCs following the treatment of ox-LDL. H19 inhibited miR-148b expression by direct interaction. Moreover, miR-148b inhibitor could reverse the effects of H19 depletion on proliferation and apoptosis in ox-LDL-stimulated HA-VSMCs. Further mechanical explorations showed that WNT1 was a target of miR-148b and H19 acted as a competing endogenous RNA (ceRNA) of miR-148b to enhance WNT1 expression. Furthermore, miR-148 inhibitor exerted its pro-proliferation and anti-apoptosis effects through activating WNT/β-catenin signaling in ox-LDL-stimulated HA-VSMCs.

Conclusion

H19 facilitated proliferation and inhibited apoptosis through modulating WNT/β-catenin signaling pathway via miR-148b in ox-LDL-stimulated HA-VSMCs, implicating the potential values of H19 in AS therapy.

NRDTD: a database for clinically or experimentally supported non-coding RNAs and drug targets associations

In recent years, more and more non-coding RNAs (ncRNAs) have been identified and increasing evidences have shown that ncRNAs may affect gene expression and disease progression, making them a new class of targets for drug discovery. It thus becomes important to understand the relationship between ncRNAs and drug targets. For this purpose, an ncRNAs and drug targets association database would be extremely beneficial. Here, we developed ncRNA Drug Targets Database (NRDTD) that collected 165 entries of clinically or experimentally supported ncRNAs as drug targets, including 97 ncRNAs and 96 drugs. Moreover, we annotated ncRNA-drug target associations with drug information from KEGG, PubChem, DrugBank, CTD or Wikipedia, GenBank sequence links, OMIM disease ID, pathway and function annotation for ncRNAs, detailed description of associations between ncRNAs and diseases from HMDD or LncRNADisease and the publication PubMed ID. Additionally, we provided users a link to submit novel disease-ncRNA-drug associations and corresponding supporting evidences into the database. We hope NRDTD will be a useful resource for investigating the roles of ncRNAs in drug target identification, drug discovery and disease treatment.

Database URL:http://chengroup.cumt.edu.cn/NRDTD

Identification and Analysis of Rice Yield-Related Candidate Genes by Walking on the Functional Network

Rice (Oryza sativa L.) is one of the most important staple foods in the world. It is possible to identify candidate genes associated with rice yield using the model of random walk with restart on a functional similarity network. We demonstrated the high performance of this approach by a five-fold cross-validation experiment, as well as the robustness of the parameter r. We also assessed the strength of associations between known seeds and candidate genes in the light of the results scores. The candidates ranking at the top of the results list were considered to be the most relevant rice yield-related genes. This study provides a valuable alternative for rice breeding and biology research. The relevant dataset and script can be downloaded at the website: http://lab.malab.cn/jj/rice.htm.

GIMDA: Graphlet interaction-based MiRNA-disease association prediction

MicroRNAs (miRNAs) have been confirmed to be closely related to various human complex diseases by many experimental studies. It is necessary and valuable to develop powerful and effective computational models to predict potential associations between miRNAs and diseases. In this work, we presented a prediction model of Graphlet Interaction for MiRNA‐Disease Association prediction (GIMDA) by integrating the disease semantic similarity, miRNA functional similarity, Gaussian interaction profile kernel similarity and the experimentally confirmed miRNA‐disease associations. The related score of a miRNA to a disease was calculated by measuring the graphlet interactions between two miRNAs or two diseases. The novelty of GIMDA lies in that we used graphlet interaction to analyse the complex relationships between two nodes in a graph. The AUCs of GIMDA in global and local leave‐one‐out cross‐validation (LOOCV) turned out to be 0.9006 and 0.8455, respectively. The average result of five‐fold cross‐validation reached to 0.8927 ± 0.0012. In case study for colon neoplasms, kidney neoplasms and prostate neoplasms based on the database of HMDD V2.0, 45, 45, 41 of the top 50 potential miRNAs predicted by GIMDA were validated by dbDEMC and miR2Disease. Additionally, in the case study of new diseases without any known associated miRNAs and the case study of predicting potential miRNA‐disease associations using HMDD V1.0, there were also high percentages of top 50 miRNAs verified by the experimental literatures.

Dihydroartemisinin-regulated mRNAs and lncRNAs in chronic myeloid leukemia

Chronic myelocytic leukemia (CML) is characterized by increased and unregulated growth of predominantly myeloid cells in the bone marrow, and accumulation of these cells in blood. We investigated the effects of an anti-malarial drug, dihydroartemisinin (DHA), on K562 CML cells. We identified 34 mRNAs and eight lncRNAs dysregulated following DHA treatment in pure and hemin-induced K562 cells. Up- or downregulation of these potential DHA targets increased with increasing DHA concentration. We also constructed and analyzed a DHA-related mRNA-lncRNA regulation network in K562 cells, and found that four DHA-modulated mRNAs regulated by four lncRNAs participated in the steroid biosynthesis pathway. Some estrogen-related drugs, such as tamoxifen, shared common targets with DHA. We inferred that DHA exerted anti-cancer effects on K562 cells by influencing estrogen levels. Our findings indicate that DHA has potential not only as an anti-malarial drug, but also as an anti-CML chemotherapeutic.

MKRMDA: multiple kernel learning-based Kronecker regularized least squares for MiRNA-disease association prediction

Background

Recently, as the research of microRNA (miRNA) continues, there are plenty of experimental evidences indicating that miRNA could be associated with various human complex diseases development and progression. Hence, it is necessary and urgent to pay more attentions to the relevant study of predicting diseases associated miRNAs, which may be helpful for effective prevention, diagnosis and treatment of human diseases. Especially, constructing computational methods to predict potential miRNA–disease associations is worthy of more studies because of the feasibility and effectivity.

Methods

In this work, we developed a novel computational model of multiple kernels learning-based Kronecker regularized least squares for MiRNA–disease association prediction (MKRMDA), which could reveal potential miRNA–disease associations by automatically optimizing the combination of multiple kernels for disease and miRNA.

Results

MKRMDA obtained AUCs of 0.9040 and 0.8446 in global and local leave-one-out cross validation, respectively. Meanwhile, MKRMDA achieved average AUCs of 0.8894 ± 0.0015 in fivefold cross validation. Furthermore, we conducted three different kinds of case studies on some important human cancers for further performance evaluation. In the case studies of colonic cancer, esophageal cancer and lymphoma based on known miRNA–disease associations in HMDDv2.0 database, 76, 94 and 88% of the corresponding top 50 predicted miRNAs were confirmed by experimental reports, respectively. In another two kinds of case studies for new diseases without any known associated miRNAs and diseases only with known associations in HMDDv1.0 database, the verified ratios of two different cancers were 88 and 94%, respectively.

Conclusions

All the results mentioned above adequately showed the reliable prediction ability of MKRMDA. We anticipated that MKRMDA could serve to facilitate further developments in the field and the follow-up investigations by biomedical researchers.

Electronic supplementary material

The online version of this article (10.1186/s12967-017-1340-3) contains supplementary material, which is available to authorized users.

PCPA protects against monocrotaline-induced pulmonary arterial remodeling in rats: potential roles of connective tissue growth factor

The purpose of this study was to investigate the mechanism of monocrotaline (MCT)-induced pulmonary artery hypertension (PAH) and determine whether 4-chloro-DL-phenylalanine (PCPA) could inhibit pulmonary arterial remodeling associated with connective tissue growth factor (CTGF) expression and downstream signal pathway. MCT was administered to forty Sprague Dawley rats to establish the PAH model. PCPA was administered at doses of 50 and 100 mg/kg once daily for 3 weeks via intraperitoneal injection. On day 22, the pulmonary arterial pressure (PAP), right ventricle hypertrophy index (RVI) and pulmonary artery morphology were assessed and the serotonin receptor-1B (SR-1B), CTGF, p-ERK/ERK were measured by western blot or immunohistochemistry. The concentration of serotonin in plasma was checked by ELISA. Apoptosis and apoptosis-related indexes were detected by TUNEL and western blot. In the MCT-induced PAH models, the PAP, RVI, pulmonary vascular remodeling, SR-1B index, CTGF index, anti-apoptotic factors bcl-xl and bcl-2, serotonin concentration in plasma were all increased and the pro-apoptotic factor caspase-3 was reduced. PCPA significantly ameliorated pulmonary arterial remodeling induced by MCT, and this action was associated with accelerated apoptosis and down-regulation of CTGF, SR-1B and p-ERK/ERK. The present study suggests that PCPA protects against the pathogenesis of PAH by suppressing remodeling and inducing apoptosis, which are likely associated with CTGF and downstream ERK signaling pathway in rats.

A novel approach based on KATZ measure to predict associations of human microbiota with non-infectious diseases

Motivation

Accumulating clinical observations have indicated that microbes living in the human body are closely associated with a wide range of human noninfectious diseases, which provides promising insights into the complex disease mechanism understanding. Predicting microbe-disease associations could not only boost human disease diagnostic and prognostic, but also improve the new drug development. However, little efforts have been attempted to understand and predict human microbe-disease associations on a large scale until now.

Results

In this work, we constructed a microbe-human disease association network and further developed a novel computational model of KATZ measure for Human Microbe-Disease Association prediction (KATZHMDA) based on the assumption that functionally similar microbes tend to have similar interaction and non-interaction patterns with noninfectious diseases, and vice versa. To our knowledge, KATZHMDA is the first tool for microbe-disease association prediction. The reliable prediction performance could be attributed to the use of KATZ measurement, and the introduction of Gaussian interaction profile kernel similarity for microbes and diseases. LOOCV and k-fold cross validation were implemented to evaluate the effectiveness of this novel computational model based on known microbe-disease associations obtained from HMDAD database. As a result, KATZHMDA achieved reliable performance with average AUCs of 0.8130 ± 0.0054, 0.8301 ± 0.0033 and 0.8382 in 2-fold and 5-fold cross validation and LOOCV framework, respectively. It is anticipated that KATZHMDA could be used to obtain more novel microbes associated with important noninfectious human diseases and therefore benefit drug discovery and human medical improvement.

Availability and Implementation

Matlab codes and dataset explored in this work are available at http://dwz.cn/4oX5mS .

Contacts

xingchen@amss.ac.cn or zhuhongyou@gmail.com or wangxuesongcumt@163.com.

Supplementary information

Supplementary data are available at Bioinformatics online.

SSCMDA: spy and super cluster strategy for MiRNA-disease association prediction

In the biological field, the identification of the associations between microRNAs (miRNAs) and diseases has been paid increasing attention as an extremely meaningful study for the clinical medicine. However, it is expensive and time-consuming to confirm miRNA-disease associations by experimental methods. Therefore, in recent years, several effective computational models for predicting the potential miRNA-disease associations have been developed. In this paper, we proposed the Spy and Super Cluster strategy for MiRNA-Disease Association prediction (SSCMDA) based on known miRNA-disease associations, integrated disease similarity and integrated miRNA similarity. For problems of mixed unknown miRNA-disease pairs containing both potential associations and real negative associations, which will lead to inaccurate prediction, spy strategy is adopted by SSCMDA to identify reliable negative samples from the unknown miRNA-disease pairs. Moreover, the super-cluster strategy could gather as many positive samples as possible to improve the accuracy of the prediction by overcoming the shortage of lacking sufficient positive training samples. As a result, the AUCs of global leave-one-out cross validation (LOOCV), local LOOCV and 5-fold cross validation were 0.9007, 0.8747 and 0.8806+/-0.0025, respectively. According to the AUC results, SSCMDA has shown a significant improvement compared with some previous models. We further carried out case studies based on various version of HMDD database to test the prediction performance robustness of SSCMDA. We also implemented case study to examine whether SSCMDA was effective for new diseases without any known associated miRNAs. As a result, a large proportion of the predicted miRNAs have been verified by experimental reports.

Graph-theoretical comparison of normal and tumor networks in identifying BRCA genes

BACKGROUND

Identification of driver genes related to certain types of cancer is an important research topic. Several systems biology approaches have been suggested, in particular for the identification of breast cancer (BRCA) related genes. Such approaches usually rely on differential gene expression and/or mutational landscape data. In some cases interaction network data is also integrated to identify cancer-related modules computationally.

RESULTS

We provide a framework for the comparative graph-theoretical analysis of networks integrating the relevant gene expression, mutations, and potein-protein interaction network data. The comparisons involve a graph-theoretical analysis of normal and tumor network pairs across all instances of a given set of breast cancer samples. The network measures under consideration are based on appropriate formulations of various centrality measures: betweenness, clustering coefficients, degree centrality, random walk distances, graph-theoretical distances, and Jaccard index centrality.

CONCLUSIONS

Among all the studied centrality-based graph-theoretical properties, we show that a betweenness-based measure differentiates BRCA genes across all normal versus tumor network pairs, than the rest of the popular centrality-based measures. The AUROC and AUPR values of the gene lists ordered with respect to the measures under study as compared to NCBI BioSystems pathway and the COSMIC database of cancer genes are the largest with the betweenness-based differentiation, followed by the measure based on degree centrality. In order to test the robustness of the suggested measures in prioritizing cancer genes, we further tested the two most promising measures, those based on betweenness and degree centralities, on randomly rewired networks. We show that both measures are quite resilient to noise in the input interaction network. We also compared the same measures against a state-of-the-art alternative disease gene prioritization method, MUFFFINN. We show that both our graph-theoretical measures outperform MUFFINN prioritizations in terms of ROC and precions/recall analysis. Finally, we filter the ordered list of the best measure, the betweenness-based differentiation, via a maximum-weight independent set formulation and investigate the top 50 genes in regards to literature verification. We show that almost all genes in the list are verified by the breast cancer literature and three genes are presented as novel genes that may potentialy be BRCA-related but missing in literature.

Genome-wide analyses of long noncoding RNA expression profiles in lung adenocarcinoma

LncRNAs have emerged as a novel class of critical regulators of cancer. We aimed to construct a landscape of lncRNAs and their potential target genes in lung adenocarcinoma. Genome-wide expression of lncRNAs and mRNAs was determined using microarray. qRT-PCR was performed to validate the expression of the selected lncRNAs in a cohort of 42 tumor tissues and adjacent normal tissues. R and Bioconductor were used for data analysis. A total of 3045 lncRNAs were differentially expressed between the paired tumor and normal tissues (1048 up and 1997 down). Meanwhile, our data showed that the expression NONHSAT077036 was associated with N classification and clinical stage. Further, we analyzed the potential co-regulatory relationship between the lncRNAs and their potential target genes using the ‘cis’ and ‘trans’ models. In the 25 related transcription factors (TFs), our analysis of The Cancer Genome Atlas database (TCGA) found that patients with lower expression of POU2F2 and higher expression of TRIM28 had a shorter overall survival time. The POU2F2 and TRIM28 co-expressed lncRNA landscape characterized here may shed light into normal biology and lung adenocarcinoma pathogenesis, and be valuable for discovery of biomarkers.

A deep ensemble model to predict miRNA-disease association

Cumulative evidence from biological experiments has confirmed that microRNAs (miRNAs) are related to many types of human diseases through different biological processes. It is anticipated that precise miRNA-disease association prediction could not only help infer potential disease-related miRNA but also boost human diagnosis and disease prevention. Considering the limitations of previous computational models, a more effective computational model needs to be implemented to predict miRNA-disease associations. In this work, we first constructed a human miRNA-miRNA similarity network utilizing miRNA-miRNA functional similarity data and heterogeneous miRNA Gaussian interaction profile kernel similarities based on the assumption that similar miRNAs with similar functions tend to be associated with similar diseases, and vice versa. Then, we constructed disease-disease similarity using disease semantic information and heterogeneous disease-related interaction data. We proposed a deep ensemble model called DeepMDA that extracts high-level features from similarity information using stacked autoencoders and then predicts miRNA-disease associations by adopting a 3-layer neural network. In addition to five-fold cross-validation, we also proposed another cross-validation method to evaluate the performance of the model. The results show that the proposed model is superior to previous methods with high robustness.

Long Noncoding RNAs as a Key Player in Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is a major malignancy in the liver and has emerged as one of the main cancers in the world with a high mortality rate. However, the molecular mechanisms of HCC are still poorly understood. Long noncoding RNAs (lncRNAs) have recently come to the forefront as functional non-protein-coding RNAs that are involved in a variety of cellular processes ranging from maintaining the structural integrity of chromosomes to gene expression regulation in a spatiotemporal manner. Many recent studies have reported the involvement of lncRNAs in HCC which has led to a better understanding of the underlying molecular mechanisms operating in HCC. Long noncoding RNAs have been shown to regulate development and progression of HCC, and thus, lncRNAs have both diagnostic and therapeutic potentials. In this review, we present an overview of the lncRNAs involved in different stages of HCC and their potential in clinical applications which have been studied so far.

LPI-NRLMF: lncRNA-protein interaction prediction by neighborhood regularized logistic matrix factorization

LncRNA-protein interactions play important roles in many important cellular processes including signaling, transcriptional regulation, and even the generation and progression of complex diseases. However, experimental methods for determining proteins bound by a specific lncRNA remain expensive, difficult and time-consuming, and only a few theoretical approaches are available for predicting potential lncRNA-protein associations. In this study, we developed a novel matrix factorization computational approach to uncover lncRNA-protein relationships, namely lncRNA-protein interactions prediction by neighborhood regularized logistic matrix factorization (LPI-NRLMF). Moreover, it is a semi-supervised and does not need negative samples. As a result, new model obtained reliable performance in the leave-one-out cross validation (the AUC of 0.9025 and AUPR of 0.6924), which significantly improved the prediction performance of previous models. Furthermore, the case study demonstrated that many lncRNA-protein interactions predicted by our method can be successfully confirmed by experiments. It is anticipated that LPI-NRLMF could serve as a useful resource for potential lncRNA-protein association identification.

PRMDA: personalized recommendation-based MiRNA-disease association prediction

Recently, researchers have been increasingly focusing on microRNAs (miRNAs) with accumulating evidence indicating that miRNAs serve as a vital role in various biological processes and dysfunctions of miRNAs are closely related with human complex diseases. Predicting potential associations between miRNAs and diseases is attached considerable significance in the domains of biology, medicine, and bioinformatics. In this study, we developed a computational model of Personalized Recommendation-based MiRNA-Disease Association prediction (PRMDA) to predict potential related miRNA for all diseases by implementing personalized recommendation-based algorithm based on integrated similarity for diseases and miRNAs. PRMDA is a global method capable of prioritizing candidate miRNAs for all diseases simultaneously. Moreover, the model could be applied to diseases without any known associated miRNAs. PRMDA obtained AUC of 0.8315 based on leave-one-out cross validation, which demonstrated that PRMDA could be regarded as a reliable tool for miRNA-disease association prediction. Besides, we implemented PRMDA on the HMDD V1.0 and HMDD V2.0 databases for three kinds of case studies about five important human cancers in order to test the performance of the model from different perspectives. As a result, 92%, 94%, 88%, 96% and 88% out of the top 50 candidate miRNAs predicted by PRMDA for Colon Neoplasms, Esophageal Neoplasms, Lymphoma, Lung Neoplasms and Breast Neoplasms, respectively, were confirmed by experimental reports.

MCMDA: Matrix completion for MiRNA-disease association prediction

Nowadays, researchers have realized that microRNAs (miRNAs) are playing a significant role in many important biological processes and they are closely connected with various complex human diseases. However, since there are too many possible miRNA-disease associations to analyze, it remains difficult to predict the potential miRNAs related to human diseases without a systematic and effective method. In this study, we developed a Matrix Completion for MiRNA-Disease Association prediction model (MCMDA) based on the known miRNA-disease associations in HMDD database. MCMDA model utilized the matrix completion algorithm to update the adjacency matrix of known miRNA-disease associations and furthermore predict the potential associations. To evaluate the performance of MCMDA, we performed leave-one-out cross validation (LOOCV) and 5-fold cross validation to compare MCMDA with three previous classical computational models (RLSMDA, HDMP, and WBSMDA). As a result, MCMDA achieved AUCs of 0.8749 in global LOOCV, 0.7718 in local LOOCV and average AUC of 0.8767+/−0.0011 in 5-fold cross validation. Moreover, the prediction results associated with colon neoplasms, kidney neoplasms, lymphoma and prostate neoplasms were verified. As a consequence, 84%, 86%, 78% and 90% of the top 50 potential miRNAs for these four diseases were respectively confirmed by recent experimental discoveries. Therefore, MCMDA model is superior to the previous models in that it improves the prediction performance although it only depends on the known miRNA-disease associations.

Identification of potential prognostic ceRNA module biomarkers in patients with pancreatic adenocarcinoma

Accumulating evidence suggested that long non-coding RNAs (lncRNAs) can function as competing endogenous RNAs (ceRNAs) to interact with other RNA transcripts and ceRNAs perturbation play important roles in cancer initiation and progression including pancreatic adenocarcinoma (PAAD). In this study, we constructed a PAAD-specific hallmark gene-related ceRNA network (HceNet) using paired genome-wide expression profiles of mRNA, lncRNA and miRNA and regulatory relationships between them. Based on “ceRNA hypothesis”, we analyzed the characteristics of HceNet and identified a ceRNA module comprising of 29 genes (12 lncRNAs, two miRNAs and 15 mRNAs) as potential prognostic biomarkers related to overall survival of patients with PAAD. The prognostic value of ceRNA module biomarkers was further validated in the train (Hazard Ratio (HR) =1.661, 95% CI: 1.275–2.165, p<1.00e-4), test (HR=1.546, 95% CI: 1.238-1.930, p<1.00e-4), and entire (HR=1.559, 95% CI: 1.321-1.839, p<1.00e-4) datasets. Our study provides candidate prognostic biomarkers for PAAD and increases our understanding of ceRNA-related regulatory mechanism in PAAD pathogenesis.

Global network random walk for predicting potential human lncRNA-disease associations

There is more and more evidence that the mutation and dysregulation of long non-coding RNA (lncRNA) are associated with numerous diseases, including cancers. However, experimental methods to identify associations between lncRNAs and diseases are expensive and time-consuming. Effective computational approaches to identify disease-related lncRNAs are in high demand; and would benefit the detection of lncRNA biomarkers for disease diagnosis, treatment, and prevention. In light of some limitations of existing computational methods, we developed a global network random walk model for predicting lncRNA-disease associations (GrwLDA) to reveal the potential associations between lncRNAs and diseases. GrwLDA is a universal network-based method and does not require negative samples. This method can be applied to a disease with no known associated lncRNA (isolated disease) and to lncRNA with no known associated disease (novel lncRNA). The leave-one-out cross validation (LOOCV) method was implemented to evaluate the predicted performance of GrwLDA. As a result, GrwLDA obtained reliable AUCs of 0.9449, 0.8562, and 0.8374 for overall, novel lncRNA and isolated disease prediction, respectively, significantly outperforming previous methods. Case studies of colon, gastric, and kidney cancers were also implemented, and the top 5 disease-lncRNA associations were reported for each disease. Interestingly, 13 (out of the 15) associations were confirmed by literature mining.

DisSetSim: an online system for calculating similarity between disease sets

Background

Functional similarity between molecules results in similar phenotypes, such as diseases. Therefore, it is an effective way to reveal the function of molecules based on their induced diseases. However, the lack of a tool for obtaining the similarity score of pair-wise disease sets (SSDS) limits this type of application.

Results

Here, we introduce DisSetSim, an online system to solve this problem in this article. Five state-of-the-art methods involving Resnik’s, Lin’s, Wang’s, PSB, and SemFunSim methods were implemented to measure the similarity score of pair-wise diseases (SSD) first. And then “pair-wise-best pairs-average” (PWBPA) method was implemented to calculated the SSDS by the SSD. The system was applied for calculating the functional similarity of miRNAs based on their induced disease sets. The results were further used to predict potential disease-miRNA relationships.

Conclusions

The high area under the receiver operating characteristic curve AUC (0.9296) based on leave-one-out cross validation shows that the PWBPA method achieves a high true positive rate and a low false positive rate. The system can be accessed from http://www.bio-annotation.cn:8080/DisSetSim/.

Gene regulation of mammalian long non-coding RNA

RNA polymerase II (Pol II) transcribes two classes of RNAs, protein-coding and non-protein-coding (ncRNA) genes. ncRNAs are also synthesized by RNA polymerases I and III (Pol I and III). In humans, the number of ncRNA genes exceeds more than twice that of protein-coding genes. However, the history of studying Pol II-synthesized ncRNA is relatively short. Since early 2000s, important biological and pathological functions of these ncRNA genes have begun to be discovered and intensively studied. And transcription mechanisms of long non-coding RNA (lncRNA) have been recently reported. Transcription of lncRNAs utilizes some transcription factors and mechanisms shared in that of protein-coding genes. In addition, tissue specificity in lncRNA gene expression has been shown. LncRNAs play essential roles in regulating the expression of neighboring or distal genes through different mechanisms. This leads to the implication of lncRNAs in a wide variety of biological pathways and pathological development. In this review, the newly discovered transcription mechanisms, characteristics, and functions of lncRNA are discussed.

Hemoglobin level, a prognostic factor for nasal extranodal natural killer/T-cell lymphoma patients from stage I to IV: A validated prognostic nomogram

Although nasal extranodal natural killer/T-cell lymphoma (nasal ENKL) shares some prognostic factors with other lymphomas, seldom studies had explored the prognostic value of hemoglobin. The ENKL cases in stage I–IV during 2000 to 2015 were collected from two medical centers (group A, n = 192), and were randomly divided into the group B (n = 155) and C (n = 37). Although the significant factors identified by the univariate analysis differed between the group A and B, the multivariate Cox regression indicated the same factors. C-index of the model was slightly better than Yang’s, but its integrated Brier score (IBS) was obviously lower than Yang’s both in the group A and B. Additionally, minimal depth of random survival forest (RSF) classifier confirmed that the prognostic ability of hemoglobin was better than age both in the group A and B. In the calibration of the nomogram, the predicted 3-year or 5-year OS of our nomogram well agreed with the corresponding actual OS. In conclusion, Hemoglobin is a prognostic factor for nasal ENKL patients in stage I - IV, and integrating it into a validated prognostic nomogram, whose generalization error is the smallest among the evaluated models, can be used to predict the patients’ outcome.

DRMDA: deep representations-based miRNA-disease association prediction

Recently, microRNAs (miRNAs) are confirmed to be important molecules within many crucial biological processes and therefore related to various complex human diseases. However, previous methods of predicting miRNA–disease associations have their own deficiencies. Under this circumstance, we developed a prediction method called deep representations‐based miRNA–disease association (DRMDA) prediction. The original miRNA–disease association data were extracted from HDMM database. Meanwhile, stacked auto‐encoder, greedy layer‐wise unsupervised pre‐training algorithm and support vector machine were implemented to predict potential associations. We compared DRMDA with five previous classical prediction models (HGIMDA, RLSMDA, HDMP, WBSMDA and RWRMDA) in global leave‐one‐out cross‐validation (LOOCV), local LOOCV and fivefold cross‐validation, respectively. The AUCs achieved by DRMDA were 0.9177, 08339 and 0.9156 ± 0.0006 in the three tests above, respectively. In further case studies, we predicted the top 50 potential miRNAs for colon neoplasms, lymphoma and prostate neoplasms, and 88%, 90% and 86% of the predicted miRNA can be verified by experimental evidence, respectively. In conclusion, DRMDA is a promising prediction method which could identify potential and novel miRNA–disease associations.

BRWLDA: bi-random walks for predicting lncRNA-disease associations

Increasing efforts have been done to figure out the association between lncRNAs and complex diseases. Many computational models construct various lncRNA similarity networks, disease similarity networks, along with known lncRNA-disease associations to infer novel associations. However, most of them neglect the structural difference between lncRNAs network and diseases network, hierarchical relationships between diseases and pattern of newly discovered associations. In this study, we developed a model that performs Bi-Random Walks to predict novel LncRNA-Disease Associations (BRWLDA in short). This model utilizes multiple heterogeneous data to construct the lncRNA functional similarity network, and Disease Ontology to construct a disease network. It then constructs a directed bi-relational network based on these two networks and available lncRNAs-disease associations. Next, it applies bi-random walks on the network to predict potential associations. BRWLDA achieves reliable and better performance than other comparing methods not only on experiment verified associations, but also on the simulated experiments with masked associations. Case studies further demonstrate the feasibility of BRWLDA in identifying new lncRNA-disease associations.

Robust and accurate prediction of protein self-interactions from amino acids sequence using evolutionary information

Self-interacting proteins (SIPs) play an essential role in cellular functions and the evolution of protein interaction networks (PINs). Due to the limitations of experimental self-interaction proteins detection technology, it is a very important task to develop a robust and accurate computational approach for SIPs prediction. In this study, we propose a novel computational method for predicting SIPs from protein amino acids sequence. Firstly, a novel feature representation scheme based on Local Binary Pattern (LBP) is developed, in which the evolutionary information, in the form of multiple sequence alignments, is taken into account. Then, by employing the Relevance Vector Machine (RVM) classifier, the performance of our proposed method is evaluated on yeast and human datasets using a five-fold cross-validation test. The experimental results show that the proposed method can achieve high accuracies of 94.82% and 97.28% on yeast and human datasets, respectively. For further assessing the performance of our method, we compared it with the state-of-the-art Support Vector Machine (SVM) classifier, and other existing methods, on the same datasets. Comparison results demonstrate that the proposed method is very promising and could provide a cost-effective alternative for predicting SIPs. In addition, to facilitate extensive studies for future proteomics research, a web server is freely available for academic use at .

Evaluation of Quality Assessment Protocols for High Throughput Genome Resequencing Data

Large-scale initiatives aiming to recover the complete sequence of thousands of human genomes are currently being undertaken worldwide, concurring to the generation of a comprehensive catalog of human genetic variation. The ultimate and most ambitious goal of human population scale genomics is the characterization of the so-called human “variome,” through the identification of causal mutations or haplotypes. Several research institutions worldwide currently use genotyping assays based on Next-Generation Sequencing (NGS) for diagnostics and clinical screenings, and the widespread application of such technologies promises major revolutions in medical science. Bioinformatic analysis of human resequencing data is one of the main factors limiting the effectiveness and general applicability of NGS for clinical studies. The requirement for multiple tools, to be combined in dedicated protocols in order to accommodate different types of data (gene panels, exomes, or whole genomes) and the high variability of the data makes difficult the establishment of a ultimate strategy of general use. While there already exist several studies comparing sensitivity and accuracy of bioinformatic pipelines for the identification of single nucleotide variants from resequencing data, little is known about the impact of quality assessment and reads pre-processing strategies. In this work we discuss major strengths and limitations of the various genome resequencing protocols are currently used in molecular diagnostics and for the discovery of novel disease-causing mutations. By taking advantage of publicly available data we devise and suggest a series of best practices for the pre-processing of the data that consistently improve the outcome of genotyping with minimal impacts on computational costs.

CarcinoPred-EL: Novel models for predicting the carcinogenicity of chemicals using molecular fingerprints and ensemble learning methods

Carcinogenicity refers to a highly toxic end point of certain chemicals, and has become an important issue in the drug development process. In this study, three novel ensemble classification models, namely Ensemble SVM, Ensemble RF, and Ensemble XGBoost, were developed to predict carcinogenicity of chemicals using seven types of molecular fingerprints and three machine learning methods based on a dataset containing 1003 diverse compounds with rat carcinogenicity. Among these three models, Ensemble XGBoost is found to be the best, giving an average accuracy of 70.1 ± 2.9%, sensitivity of 67.0 ± 5.0%, and specificity of 73.1 ± 4.4% in five-fold cross-validation and an accuracy of 70.0%, sensitivity of 65.2%, and specificity of 76.5% in external validation. In comparison with some recent methods, the ensemble models outperform some machine learning-based approaches and yield equal accuracy and higher specificity but lower sensitivity than rule-based expert systems. It is also found that the ensemble models could be further improved if more data were available. As an application, the ensemble models are employed to discover potential carcinogens in the DrugBank database. The results indicate that the proposed models are helpful in predicting the carcinogenicity of chemicals. A web server called CarcinoPred-EL has been built for these models (http://ccsipb.lnu.edu.cn/toxicity/CarcinoPred-EL/).

SASP regulation by noncoding RNA

Noncoding RNAs (ncRNAs), including micro (mi)RNAs, long noncoding (lnc)RNAs, and circular (circ)RNAs, control specific gene expression programs by regulating transcriptional, post-transcriptional, and post-translational processes. Through their broad influence on protein expression and function, ncRNAs have been implicated in virtually all cellular processes such as proliferation, senescence, quiescence, differentiation, apoptosis, and the stress and immune responses. Senescence is a cellular phenotype associated with the physiologic decline of aging and with age-related pathologies. Besides their characteristic terminal growth arrest and differential gene expression programs, senescent cells are known to secrete potent pro-inflammatory, angiogenic, and tissue-remodeling factors. This important trait, known as the senescence-associated secretory phenotype (SASP), influences many biological processes such as tissue repair and regeneration, tumorigenesis, and the aging-associated pro-inflammatory state. Here, we review the microRNAs, lncRNAs, and circRNAs that influence the production of SASP factors and discuss the rising interest in SASP-regulatory ncRNAs as diagnostic and therapeutic targets.

Autophagy-related gene expression is an independent prognostic indicator of glioma

In this study, we identified 74 differentially expressed autophagy-related genes in glioma patients. Analysis using a Cox proportional hazard regression model showed that MAPK8IP1 and SH3GLB1, two autophagy-related genes, were associated with the prognostic signature for glioma. Glioma patients from the CGGA batches 1 and 2, GSE4412 and TCGA datasets could be divided into high- and low-risk groups with different survival times based on levels of MAPK8IP1 and SH3GLB1 expression. The autophagy-related signature was an independent predictor of survival outcomes in glioma patients. MAPK8IP1 overexpression and SH3GLB1 knockdown inhibited glioma cell proliferation, migration and invasion, and improved Temozolomide sensitivity. These findings suggest autophagy-related genes like MAPK8IP1 and SH3GLB1 could be potential therapeutic targets in glioma.

A Review on Recent Computational Methods for Predicting Noncoding RNAs

Noncoding RNAs (ncRNAs) play important roles in various cellular activities and diseases. In this paper, we presented a comprehensive review on computational methods for ncRNA prediction, which are generally grouped into four categories: (1) homology-based methods, that is, comparative methods involving evolutionarily conserved RNA sequences and structures, (2) de novo methods using RNA sequence and structure features, (3) transcriptional sequencing and assembling based methods, that is, methods designed for single and pair-ended reads generated from next-generation RNA sequencing, and (4) RNA family specific methods, for example, methods specific for microRNAs and long noncoding RNAs. In the end, we summarized the advantages and limitations of these methods and pointed out a few possible future directions for ncRNA prediction. In conclusion, many computational methods have been demonstrated to be effective in predicting ncRNAs for further experimental validation. They are critical in reducing the huge number of potential ncRNAs and pointing the community to high confidence candidates. In the future, high efficient mapping technology and more intrinsic sequence features (e.g., motif and k-mer frequencies) and structure features (e.g., minimum free energy, conserved stem-loop, or graph structures) are suggested to be combined with the next- and third-generation sequencing platforms to improve ncRNA prediction.

A five-long non-coding RNA signature to improve prognosis prediction of clear cell renal cell carcinoma

Recent works have reported that long non-coding RNAs (lncRNAs) play critical roles in tumorigenesis and prognosis of cancers, suggesting the potential utility of lncRNAs as cancer prognostic markers. However, lncRNA signatures in predicting the survival of patients with clear cell renal cell carcinoma (ccRCC) remain unknown. In this study, we attempted to identify lncRNA signatures and their prognostic values in ccRCC. Using lncRNA expression profiling data in 440 ccRCC tumors from The Cancer Genome Atlas (TCGA) data, a five-lncRNA signature (AC069513.4, AC003092.1, CTC-205M6.2, RP11-507K2.3, U91328.21) has been identified to be significantly associated with ccRCC patients’ overall survival in both training set and testing set. Based on the lncRNA signature, ccRCC patients could be divided into high-risk and low-risk group with significantly different survival rate. Further multivariable Cox regression analysis suggested that the prognostic value of this signature was independent of clinical factors. Functional enrichment analyses showed the potential functional roles of the five prognostic lncRNAs in ccRCC oncogenesis. These results indicated that this five-lncRNA signature could be used as an independent prognostic biomarker in the prediction of ccRCC patients’ survival.

RKNNMDA: Ranking-based KNN for MiRNA-Disease Association prediction

Cumulative verified experimental studies have demonstrated that microRNAs (miRNAs) could be closely related with the development and progression of human complex diseases. Based on the assumption that functional similar miRNAs may have a strong correlation with phenotypically similar diseases and vice versa, researchers developed various effective computational models which combine heterogeneous biologic data sets including disease similarity network, miRNA similarity network, and known disease-miRNA association network to identify potential relationships between miRNAs and diseases in biomedical research. Considering the limitations in previous computational study, we introduced a novel computational method of Ranking-based KNN for miRNA-Disease Association prediction (RKNNMDA) to predict potential related miRNAs for diseases, and our method obtained an AUC of 0.8221 based on leave-one-out cross validation. In addition, RKNNMDA was applied to 3 kinds of important human cancers for further performance evaluation. The results showed that 96%, 80% and 94% of predicted top 50 potential related miRNAs for Colon Neoplasms, Esophageal Neoplasms, and Prostate Neoplasms have been confirmed by experimental literatures, respectively. Moreover, RKNNMDA could be used to predict potential miRNAs for diseases without any known miRNAs, and it is anticipated that RKNNMDA would be of great use for novel miRNA-disease association identification.

Computational prediction of human disease-related microRNAs by path-based random walk

MicroRNAs (miRNAs) are a class of small, endogenous RNAs that are 21–25 nucleotides in length. In animals and plants, miRNAs target specific genes for degradation or translation repression. Discovering disease-related miRNA is fundamental for understanding the pathogenesis of diseases. The association between miRNA and a disease is mainly determined via biological investigation, which is complicated by increased biological information due to big data from different databases. Researchers have utilized different computational methods to harmonize experimental approaches to discover miRNA that articulates restrictively in specific environmental situations. In this work, we present a prediction model that is based on the theory of path features and random walk to obtain a relevancy score of miRNA-related disease. In this model, highly ranked scores are potential miRNA-disease associations. Features were extracted from positive and negative samples of miRNA-disease association. Then, we compared our method with other presented models using the five-fold cross-validation method, which obtained an area under the receiver operating characteristic curve of 88.6%. This indicated that our method has a better performance compared to previous methods and will help future biological investigations.

Meta-signature LncRNAs serve as novel biomarkers for colorectal cancer: integrated bioinformatics analysis, experimental validation and diagnostic evaluation

The aim of this study is to explore the differentially expressed lncRNAs, which may have potential biological function and diagnostic value in colorectal cancer (CRC). Through integrated data mining, we finally identified nine differentially expressed lncRNAs and their potential mRNA targets. After a series of bioinformatics analyses, we screened significant pathways and GO terms that are related to the up-regulated and down-regulated transcripts respectively. Meanwhile, the nine lncRNAs were validated in 30 paired tissues and cell lines by qRT-PCR and the results were basically consistent with the microarray data. We also tested the nine lncRNAs in the serum of 30 CRC patients matched with the CRC tissue, 30 non-cancer patients and 30 health controls. Finally, we found that BLACAT1 was significant for the diagnosis of CRC. The area under the curve (AUC), sensitivity and specificity were 0.858 (95% CI: 0.765-0.951), 83.3% and 76.7% respectively between CRC patients and health controls. Moreover, BLACAT1 also had distinct value to discriminate CRC from other non-cancer diseases. The results indicated that the differentially expressed lncRNAs and their potential target transcripts could be considered as potential therapeutic targets for CRC patients. Meanwhile, lncRNA BLACAT1 might represent a new supplementary biomarker for the diagnosis of CRC.

The Evolution and Expression Pattern of Human Overlapping lncRNA and Protein-coding Gene Pairs

Long non-coding RNA overlapping with protein-coding gene (lncRNA-coding pair) is a special type of overlapping genes. Protein-coding overlapping genes have been well studied and increasing attention has been paid to lncRNAs. By studying lncRNA-coding pairs in human genome, we showed that lncRNA-coding pairs were more likely to be generated by overprinting and retaining genes in lncRNA-coding pairs were given higher priority than non-overlapping genes. Besides, the preference of overlapping configurations preserved during evolution was based on the origin of lncRNA-coding pairs. Further investigations showed that lncRNAs promoting the splicing of their embedded protein-coding partners was a unilateral interaction, but the existence of overlapping partners improving the gene expression was bidirectional and the effect was decreased with the increased evolutionary age of genes. Additionally, the expression of lncRNA-coding pairs showed an overall positive correlation and the expression correlation was associated with their overlapping configurations, local genomic environment and evolutionary age of genes. Comparison of the expression correlation of lncRNA-coding pairs between normal and cancer samples found that the lineage-specific pairs including old protein-coding genes may play an important role in tumorigenesis. This work presents a systematically comprehensive understanding of the evolution and the expression pattern of human lncRNA-coding pairs.

PBMDA: A novel and effective path-based computational model for miRNA-disease association prediction

In the recent few years, an increasing number of studies have shown that microRNAs (miRNAs) play critical roles in many fundamental and important biological processes. As one of pathogenetic factors, the molecular mechanisms underlying human complex diseases still have not been completely understood from the perspective of miRNA. Predicting potential miRNA-disease associations makes important contributions to understanding the pathogenesis of diseases, developing new drugs, and formulating individualized diagnosis and treatment for diverse human complex diseases. Instead of only depending on expensive and time-consuming biological experiments, computational prediction models are effective by predicting potential miRNA-disease associations, prioritizing candidate miRNAs for the investigated diseases, and selecting those miRNAs with higher association probabilities for further experimental validation. In this study, Path-Based MiRNA-Disease Association (PBMDA) prediction model was proposed by integrating known human miRNA-disease associations, miRNA functional similarity, disease semantic similarity, and Gaussian interaction profile kernel similarity for miRNAs and diseases. This model constructed a heterogeneous graph consisting of three interlinked sub-graphs and further adopted depth-first search algorithm to infer potential miRNA-disease associations. As a result, PBMDA achieved reliable performance in the frameworks of both local and global LOOCV (AUCs of 0.8341 and 0.9169, respectively) and 5-fold cross validation (average AUC of 0.9172). In the cases studies of three important human diseases, 88% (Esophageal Neoplasms), 88% (Kidney Neoplasms) and 90% (Colon Neoplasms) of top-50 predicted miRNAs have been manually confirmed by previous experimental reports from literatures. Through the comparison performance between PBMDA and other previous models in case studies, the reliable performance also demonstrates that PBMDA could serve as a powerful computational tool to accelerate the identification of disease-miRNA associations.

Identification of miRNA-disease associations is considered as a key way for the development of pathology, diagnose and therapy. Computational prediction models contribute to discovering the underlying disease-related miRNAs on a large scale. Based on the assumption that functionally related miRNAs tend to be involved in phenotypically similar disease and vice versa, the model of PBMDA was developed to prioritize the underlying miRNA-disease associations by adopting a special depth-first search algorithm in a heterogeneous graph, which was composed of known miRNA-disease association network, miRNA similarity network, and disease similarity network. Through leave-one-out cross validation and 5-fold cross validation, the promising results demonstrated the effectiveness of the proposed model. We further implemented the case studies of three important human complex diseases, 88%, 88% and 90% of top-50 predicted miRNA-disease associations have been manually confirmed based on recent experimental reports. It is anticipated that PBMDA could prioritize the most potential miRNA-disease associations on a large scale for advancing the progress of biological experiment validation in the future, which could further contribute to the understanding of complex disease mechanisms.

Long non-coding RNA 00312 regulated by HOXA5 inhibits tumour proliferation and promotes apoptosis in Non-small cell lung cancer

Non‐small cell lung cancer (NSCLC) is the most prevalent type of lung cancer. The abnormal expression of many long non‐coding RNAs (lncRNAs) has been reported involved in the progression of various tumours, which can be used as diagnostic indicators or antitumour targets. Here, we found that the long non‐coding RNA 00312 was down‐regulated in paired NSCLC tissues and correlated with poor clinical outcome; decreased linc00312 expression in NSCLC was associated with larger and later stage tumours. Functional experiments showed that linc00312 could inhibit cell proliferation and promote apoptosis in vitro and in vivo. Furthermore, we found that HOXA5 could bind in the promoter of linc00312 and up‐regulated the expression of it. Moreover, linc00312 was down‐regulated in the plasma of NSCLC patients compared with that of healthy volunteers or other pulmonary diseases patients. Taken together, our findings indicated that linc00312 could be a novel diagnosis biomarker and a promising therapeutic target for NSCLC.

Seven LncRNA-mRNA based risk score predicts the survival of head and neck squamous cell carcinoma

Dysregulation of mRNAs and long non-coding RNAs (lncRNAs) is one of the most important features of carcinogenesis and cancer development. However, studies integrating the expression of mRNAs and lncRNAs to predict the survival of head and neck squamous cell carcinoma (HNSC) are still limited, hitherto. In current work, we identified survival related mRNAs and lncRNAs in three datasets (TCGA dataset, E-TABM-302, GSE41613). By random forest, seven gene signatures (six mRNAs and lncRNA) were further selected to develop the risk score model. The risk score was significantly associated with survival in both training and testing datasets (E-TABM-302, GSE41613, and E-MTAB-1324). Furthermore, correlation analyses showed that the risk score is independent from clinicopathological features. According to Cox multivariable hazard model and nomogram, the risk score contributes the most to survival than the other clinical information, including gender, age, histologic grade, and alcohol taking. The Gene Set Enrichment Analysis (GSEA) indicates that the risk score is associated with cancer related pathways. In summary, the lncRNA-mRNA based risk score model we developed successfully predicts the survival of 755 HNSC samples in five datasets and two platforms. It is independent from clinical information and performs better than clinical information for prognosis.