Computational approaches to discovering noncoding RNA

Comparison and benchmark of deep learning methods for non-coding RNA classification

The involvement of non-coding RNAs in biological processes and diseases has made the exploration of their functions crucial. Most non-coding RNAs have yet to be studied, creating the need for methods that can rapidly classify large sets of non-coding RNAs into functional groups, or classes. In recent years, the success of deep learning in various domains led to its application to non-coding RNA classification. Multiple novel architectures have been developed, but these advancements are not covered by current literature reviews. We present an exhaustive comparison of the different methods proposed in the state-of-the-art and describe their associated datasets. Moreover, the literature lacks objective benchmarks. We perform experiments to fairly evaluate the performance of various tools for non-coding RNA classification on popular datasets. The robustness of methods to non-functional sequences and sequence boundary noise is explored. We also measure computation time and CO2 emissions. With regard to these results, we assess the relevance of the different architectural choices and provide recommendations to consider in future methods.

Screening and evaluation of key genes in contributing to pathogenesis of hepatic fibrosis based on microarray data

BACKGROUND

Hepatic fibrosis (HF), which is characterized by the excessive accumulation of extracellular matrix (ECM) in the liver, usually progresses to liver cirrhosis and then death. To screen differentially expressed (DE) long non-coding RNAs (lncRNAs) and mRNAs, explore their potential functions to elucidate the underlying mechanisms of HF.

METHODS

The microarray of GSE80601 was downloaded from the Gene Expression Omnibus database, which is based on the GPL1355 platform. Screening for the differentially expressed LncRNAs and mRNAs was conducted between the control and model groups. Then, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to analyze the biological functions and pathways of the DE mRNAs. Additionally, the protein-protein interaction (PPI) network was delineated. In addition, utilizing the Weighted Gene Co-expression Network Analysis (WGCNA) package and Cytoscape software, we constructed lncRNA-mRNA weighted co-expression networks.

RESULTS

A total of 254 significantly differentially expressed lncRNAs and 472 mRNAs were identified. GO and KEGG analyses revealed that DE mRNAs regulated HF by participating in the GO terms of metabolic process, inflammatory response, response to wounding and oxidation-reduction. DE mRNAs were also significantly enriched in the pathways of ECM-receptor interaction, PI3K-Akt signaling pathway, focal adhesion (FA), retinol metabolism and metabolic pathways. Moreover, 24 lncRNAs associated with 40 differentially expressed genes were observed in the modules of lncRNA-mRNA weighted co-expression network.

CONCLUSIONS

This study revealed crucial information on the molecular mechanisms of HF and laid a foundation for subsequent genes validation and functional studies, which could contribute to the development of novel diagnostic markers and provide new therapeutic targets for the clinical treatment of HF.

Key Anti-Fibrosis Associated Long Noncoding RNAs Identified in Human Hepatic Stellate Cell via Transcriptome Sequencing Analysis

Hepatic fibrosis is the main pathological basis for chronic cirrhosis, and activated hepatic stellate cells (HSCs) are the primary cells involved in liver fibrosis. Our study analyzed anti-fibrosis long noncoding RNAs (lncRNAs) in activated human HSCs (hHSCs). We performed RNA sequencing (RNA-seq) and bioinformatics analysis to determine whether lncRNA expression profile changes between hHSCs activation and quiescence. Eight differentially expressed (DE) lncRNAs and three pairs of co-expression lncRNAs-mRNAs were verified by quantitative Real-Time Polymerase Chain Reaction (qRT-PCR). A total of 34146 DE lncRNAs were identified in this study. Via gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses, we found several DE lncRNAs regulated hHSC activation by participating in DNA bending/packaging complex, growth factor binding and the Hippo signaling pathway (p < 0.05). With lncRNA–mRNA co-expression analysis, three lncRNAs were identified to be associated with connective tissue growth factor (CTGF), fibroblast growth factor 2 (FGF2) and netrin-4 (NTN4). The quantitative Real-Time Polymerase Chain Reaction (qRT-PCR) results of the eight DE lncRNAs and three pairs of co-expression lncRNAs–mRNAs were consistent with the RNA-seq data and previous reports. Several lncRNAs may serve as potential targets to reverse the progression of liver fibrosis. This study provides a first insight into lncRNA expression profile changes associated with activated human HSCs.

Long non-coding RNA H1 promotes cell proliferation and invasion by acting as a ceRNA of miR‑138 and releasing EZH2 in oral squamous cell carcinoma

Long non-coding RNAs (lncRNAs) have been shown to play pivotal roles in various types of human cancer, including oral squamous cell carcinoma (OSCC). However, the potential mechanisms of action of lncRNAs in OSCC remain to be fully elucidated. The aim of the present study was to further explore the potential mechanisms of action of lncRNAs in OSCC. We first analyzed Gene Expression Omnibus (GEO) datasets to investigate aberrantly expressed lncRNAs which may be involved in the development of OSCC. Reverse transcription‑quantitative PCR (RT‑qPCR) was performed to analyze the expression levels of lncRNA H19. In addition, the correlation between H19 expression and the clinical characteristics and prognosis of patients with OSCC was statistically analyzed. The effects of H19 expression on OSCC cells were examined by using overexpression and RNA interference approaches in vitro and in vivo. To examine the competitive endogenous RNA (ceRNA) mechanisms, bioinformatics analysis and luciferase reporter assay were performed. In addition, the correlation between H19 and microRNA (miR)‑138 was detected. H19 was found to be upregulated in OSCC tissues and its high expression level was associated with the TNM stage and nodal invasion, and also correlated with a shorter overall survival of patients with OSCC. The knockdown of H19 significantly inhibited OSCC cell proliferation, migration, invasion and epithelial-mesenchymal transition (EMT), and induced apoptosis in vitro; it also suppressed subcutaneous tumor growth in vivo. In addition, H19 was found to regulate the expression of oncogene enhancer of zeste homolog 2 (EZH2) by competing with miR‑138; the inhibition of miR‑138 attenuated the inhibitory effects of H19 knockdown on OSCC cells. On the whole, our findings suggest that H19 functions as an oncogene by inhibiting miR‑138 and facilitating EZH2 expression in OSCC. Thus, lncRNA H1 may represent a potential therapeutic target for OSCC.

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.

Long non-coding RNA ATB promotes glioma malignancy by negatively regulating miR-200a

Background

Glioma is one of the most common and aggressive primary malignant tumor in the brain. Accumulating evidences indicated that aberrantly expressed non-coding RNAs (ncRNAs), including long non-coding RNAs (lncRNAs) and microRNAs (miRNAs), contribute to tumorigenesis. However, potential mechanisms between lncRNAs and miRNAs in glioma remain largely unknown.

Methods

Long non-coding RNA activated by TGF-β (LncRNA-ATB) expression in glioma tissues and cells was quantified by quantitative reverse transcription–PCR. Glioma cell lines U251 and A172 were transfected with sh-ATB, miR-200a mimics, miR-200a inhibitors, after we assayed the cell phenotype and expression of the relevant molecules. Dual-luciferase reporter assay, RIP and a xenograft mouse model were used to examine the expression of sh-ATB and its target gene miR-200a.

Results

ATB is abnormally up-regulated both in glioma tissues and cell lines compared with normal brain tissues, and glioma patients with high ATB expression had shorter overall survival time. Knockdown of ATB significantly inhibits glioma malignancy, including cell proliferation, colony formation, migration, invasion in vitro, and the xenograft tumor formation in vivo. In addition, ATB was confirmed to target miR-200a, and miR-200a inhibition reversed the malignant characteristics of ATB knockdown on glioma cells. In particular, ATB may act as a ceRNA, effectively becoming a sink for miR-200a, thereby modulating the derepression of TGF-β2.

Conclusions

Our findings suggest that ATB plays an oncogenic role of glioma cells by inhibiting miR-200a and facilitating TGF-β2 in glioma, thereby may represent a potential therapeutic target for the treatment of human glioma.

The Underexploited Role of Non-Coding RNAs in Lysosomal Storage Diseases

Non-coding RNAs (ncRNAs) are a functional class of RNA involved in the regulation of several cellular processes which may modulate disease onset, progression, and prognosis. Lysosomal storage diseases (LSD) are a group of rare disorders caused by mutations of genes encoding specific hydrolases or non-enzymatic proteins, characterized by a wide spectrum of manifestations. The alteration of ncRNA levels is well established in several human diseases such as cancer and auto-immune disorders; however, there is a lack of information focused on the role of ncRNA in rare diseases. Recent reports related to changes in ncRNA expression and its consequences on LSD physiopathology show us the importance to keep advancing in this field. This article will summarize recent findings and provide key points for further studies on LSD and ncRNA association.

NEAT expression is associated with tumor recurrence and unfavorable prognosis in colorectal cancer

Long noncoding RNAs (lncRNAs) have recently been identified to be involved in various diseases including cancer. NEAT1 is a recently identified lncRNA with its function largely unknown in human malignancy. In the present study, we investigated NEAT1 expression in 239 cases of clinical colorectal cancer specimens and matched normal tissues. Statistical methods were utilized to analyze the association of NEAT1 with clinical features, disease-free and overall survival of patients. Results showed that NEAT1 expression in colorectal cancer was up-regulated in 72.0% (172/239) cases compared with corresponding normal counterparts, and related to tumor differentiation, invasion, metastasis and TNM stage. Kaplan-Meier analysis proved that NEAT1 was associated with both disease-free survival and overall survival of patients with colorectal cancer that patients with high NEAT1 expression tend to have unfavorable outcome. Moreover, cox's proportional hazards analysis showed that high NEAT1 expression was an independent prognostic marker of poor outcome. These results provided the first evidence that the expression of NEAT1 in colorectal cancer may play an oncogenic role in colorectal cancer differentiation, invasion and metastasis. It also proved that NEAT1 may serve as an indicator of tumor recurrence and prognosis of colorectal cancer.

Long noncoding RNA-LET, which is repressed by EZH2, inhibits cell proliferation and induces apoptosis of nasopharyngeal carcinoma cell

Recent studies have reported that long noncoding RNAs (lncRNAs) play critical roles in carcinogenesis and progression. LncRNA-LET, a recently identified lncRNA, has been shown to be a tumor suppressor in hepatocellular carcinoma. However, the expression and functional of lncRNA-LET in other type of cancers remain largely unknown. In this study, we found that lncRNA-LET was significantly downregulated in nasopharyngeal carcinoma (NPC) tissues compared with corresponding normal tissues. Decreased LET expression is significantly correlated with advanced clinical stage, larger tumor size, increased lymph node tumor burden, and poor survival of NPC patients. Gain- and loss-of-function experiments demonstrated that enhanced LET expression inhibited NPC cells proliferation and induced cell apoptosis. By contrast, the knockdown of LET promoted NPC cells proliferation and inhibited cell apoptosis. Importantly, we found lncRNA-LET is transcriptional repressed by EZH2-mediated H3K27 histone methylation on the LET promoter. The expressions of EZH2 and lncRNA-LET are significantly inversely correlated in NPC tissues. Collectively, these findings indicate a pivotal role for lncRNA-LET in NPC cell proliferation and apoptosis, and reveal an epigenetic mechanism for lncRNA-LET dysregulation.

Downregulation of a long noncoding RNA-ncRuPAR contributes to tumor inhibition in colorectal cancer

In recent years, the role of long noncoding RNAs (lncRNAs) in cancer is increasingly focused. ncRuPAR is a newly detected lncRNA; in previous study, we found out that ncRuPAR could inhibit tumor progression by downregulating protease-activated receptor-1 (PAR-1), but its role in colorectal cancer (CRC) is never elucidated. Here, we conducted a self-control study which includes 105 CRC samples. By quantitative real time PCR (qRT-PCR) and immunohistochemical staining, we detected the expression of ncRuPAR and PAR-1 as well as their correlation; we further associated these data with the clinicopathologic parameters. A receiver operating characteristic (ROC) curve was constructed to evaluate the diagnostic value of ncRuPAR and PAR-1, respectively. Our results indicated that the expression of ncRuPAR was significantly downregulated in CRC compared with paired adjacent nontumor tissues, but the level of PAR-1 mRNA in cancerous tissues was significantly higher than in adjacent normal areas. The expression of ncRuPAR was significantly correlated with lymph node metastasis, distant metastasis, Duck's stage, differentiation, and TNM stage and was potentially negatively associated with the mRNA levels and EI scores of PAR-1. The area under the ROC curve of ncRuPAR was 0.81 (95% confidence interval (CI): 0.75-0.87); at a cutoff value of 8.34, the ncRuPAR measurement had a sensitivity of 97.14%, a specificity of 65.87%, and an accuracy of 82.86% to predict CRC.

Fast selection of miRNA candidates based on large-scale pre-computed MFE sets of randomized sequences

Background

Small RNAs are important regulators of genome function, yet their prediction in genomes is still a major computational challenge. Statistical analyses of pre-miRNA sequences indicated that their 2D structure tends to have a minimal free energy (MFE) significantly lower than MFE values of equivalently randomized sequences with the same nucleotide composition, in contrast to other classes of non-coding RNA. The computation of many MFEs is, however, too intensive to allow for genome-wide screenings.

Results

Using a local grid infrastructure, MFE distributions of random sequences were pre-calculated on a large scale. These distributions follow a normal distribution and can be used to determine the MFE distribution for any given sequence composition by interpolation. It allows on-the-fly calculation of the normal distribution for any candidate sequence composition.

Conclusion

The speedup achieved makes genome-wide screening with this characteristic of a pre-miRNA sequence practical. Although this particular property alone will not be able to distinguish miRNAs from other sequences sufficiently discriminative, the MFE-based P-value should be added to the parameters of choice to be included in the selection of potential miRNA candidates for experimental verification.

Long noncoding RNA CCAT1, which could be activated by c-Myc, promotes the progression of gastric carcinoma

PURPOSE

Long noncoding RNAs (lncRNAs) have been shown to have functional roles in cancer biology and are deregulated in many tumors. The specific aim of this study was to determine the role of a long noncoding RNA CCAT1 in the progression of gastric carcinoma and discover which factors contribute to the deregulation of CCAT1.

METHODS

A computational screen of CCAT1 promoter was conducted to search for transcription-factor-binding sites. The association of c-Myc with the CCAT1 promoter in vivo was tested by chromatin immunoprecipitation assay. CCAT1 promoter activities were examined by luciferase reporter assay. The function of the c-Myc binding site in the CCAT1 promoter region was tested by a promoter assay with nucleotide substitutions in the putative E-box. The effect of CCAT1 on gastric carcinoma cell proliferation and migration was tested using in vitro cell proliferation and migration assays.

RESULTS

CCAT1 levels were markedly increased in gastric carcinoma tissues compared with normal tissues. c-Myc directly binds to the E-box element in the promoter region of CCAT1, and when ectopically expressed increased promoter activity and expression of CCAT1. Nucleotide substitutions in the E-box element in the promoter region abrogated c-Myc-dependent promoter activation. The expression of CCAT1 and c-Myc shows strong association in gastric carcinomas. Moreover, abnormally expressed CCAT1 promotes cell proliferation and migration.

CONCLUSIONS

These data suggest that c-Myc induction of CCAT1 holds an important role in gastric carcinoma and implicate the potential application of CCAT1 in the treatment of gastric carcinoma.

Panning for Long Noncoding RNAs

The recent advent of high-throughput approaches has revealed widespread transcription of the human genome, leading to a new appreciation of transcription regulation, especially from noncoding regions. Distinct from most coding and small noncoding RNAs, long noncoding RNAs (lncRNAs) are generally expressed at low levels, are less conserved and lack protein-coding capacity. These intrinsic features of lncRNAs have not only hampered their full annotation in the past several years, but have also generated controversy concerning whether many or most of these lncRNAs are simply the result of transcriptional noise. Here, we assess these intrinsic features that have challenged lncRNA discovery and further summarize recent progress in lncRNA discovery with integrated methodologies, from which new lessons and insights can be derived to achieve better characterization of lncRNA expression regulation. Full annotation of lncRNA repertoires and the implications of such annotation will provide a fundamental basis for comprehensive understanding of pervasive functions of lncRNAs in biological regulation.

A long non-coding RNA, PTCSC3, as a tumor suppressor and a target of miRNAs in thyroid cancer cells

Papillary thyroid carcinoma susceptibility candidate 3 (PTCSC3) is a newly identified non-coding RNA, which is highly thyroid-specific. Dramatic downregulation in thyroid cancers suggests its potential roles in the occurrence and development of thyroid tumors. The present study aimed to investigate the effects of PTCSC3 on the biological features of thyroid cancer cells and to explore its possible function as a competing endogenous RNA to bind with miRNAs. Constructs containing the long non-coding RNA, PTCSC3, were transfected into various thyroid cancer cell lines (BCPAP, FTC133 and 8505C). Cell growth, cell cycle transition and apoptosis were measured by MTT assay and flow cytometry. In silico analysis was performed to identify the binding site of PTCSC3 for target miRNAs. Additionally, detection of putative miRNA by quantitative reverse transcription-polymerase chain reaction (RT-PCR) in thyroid cancer cells transfected with PTCSC3 was determined to confirm the interaction. Following transfection with PTCSC3, all three thyroid cancer cells originating from various pathological types of thyroid cancers demonstrated significant growth inhibition, cell cycle arrest and increased apoptosis. The top 20 miRNAs to have a potential interaction with PTCSC3 were identified, out of which miR-574-5p was selected to further confirm the inverse correlation with PTCSC3 in thyroid cancer cells in vitro. In the present study, PTCSC3 as a tumor suppressor was investigated as a competing endogenous RNA for miR-574-5p.