Decoding the Non-coding: Tools and Databases Unveiling the Hidden World of “Junk” RNAs for Innovative Therapeutic Exploration

Abstract

Non-coding RNAs are pivotal regulators of gene and protein expression, exerting crucial influences on diverse biological processes. Their dysregulation is frequently implicated in the onset and progression of diseases, notably cancer. A profound comprehension of the intricate mechanisms governing ncRNAs is imperative for devising innovative therapeutic interventions against these debilitating conditions. Significantly, nearly 80% of our genome comprises ncRNAs, underscoring their centrality in cellular processes. The elucidation of ncRNA functions is pivotal for grasping the complexities of gene regulation and its implications for human health. Modern genome sequencing techniques yield vast datasets, stored in specialized databases. To harness this wealth of information and to understand the crosstalk of non-coding RNAs, knowledge of available databases is required, and many new sophisticated computational tools have emerged. These tools play a pivotal role in the identification, prediction, and annotation of ncRNAs, thereby facilitating their experimental validation. This Review succinctly outlines the current understanding of ncRNAs, emphasizing their involvement in disease development. It also highlights the databases and tools instrumental in classifying, annotating, and evaluating ncRNAs. By extracting meaningful biological insights from seemingly “junk” data, these tools empower scientists to unravel the intricate roles of ncRNAs in shaping human health.

Sequencing the human genome unveiled new dimensions in molecular mechanisms, with a small portion translating into proteins, <2% of the genome. The non-coding (nc) genome’s functions remain unsolved, but recent studies highlight ncRNAs’ crucial roles in diverse biological processes and disease pathogenesis. Sequencing and analysis revealed 98% of the transcriptional activity in “Junk DNA”, prompting questions about gene evolution of an independent gene type or a convergence of coding and non-coding genes.¹ Non-coding is further divided into “housekeeping ncRNAs” and “regulatory ncRNAs”, and then regulatory ncRNAs are categorized based on size into “small ncRNAs” (<200nt) and “large ncRNAs” (>200nt), as illustrated in Figure 1.

Figure 1.

Overview of different types of ncRNAs.

There are different kinds of ncRNAs, such as rRNA, snRNA, tRNA, microRNA, lncRNA, circRNA, and piwiRNA. A plethora of novel transcripts determined by advanced deep sequencing technologies required the classification and annotation of ncRNAs. The HUGO Gene Nomenclature Committee (HGNC) provides standard nomenclature of ncRNA genes annotated from the human genome.² This nomenclature of ncRNA genes ensures that they are correctly cited and prevents confusion of gene symbols. If the ncRNAs do not have gene symbols or are not classified, it would create confusion for researchers while annotating the data. Non-coding RNAs are briefly explained as follows: First, lncRNAs are transcribed by RNA POL II. LncRNAs are epigenetic regulators of functions in histone and DNA modification, primarily methylation and acetylation, to control the transcription of genes.³ Classification of lncRNAs is based on length, location with respect to protein-coding genes, residence within specific DNA regulatory elements and loci, biogenesis pathway, subcellular localization or origin, lncRNA function, or association with specific biological processes.⁴Second, biogenesis of microRNA begins from polycistronic transcripts found in an intron or UTR of protein coding genes (PCGs). MicroRNA plays a critical regulatory role in inhibiting protein translation and destabilizing mRNA and thereby deregulating post-transcriptional gene regulation.⁵ MicroRNAs are classified on the basis of miRNA associated with disease, miRNA associated with function, miRNA associated with cluster genes, and miRNA with tissue specificity. Third, circRNA, produced from back-splicing, controls genes by inhibiting miRNA activity as a miRNA sponge, and it control one or more miRNAs.⁶ CircRNA can be classified into four types depending on its origin: (i) exonic circRNAs (ecircRNAs), which consist exclusively of single or multiple exons; (ii) circular intronic RNAs (ciRNAs), which are intron-derived; (iii) exon-intron circRNAs (ElciRNAs), which contain both exons and introns; and (iv) tRNA intronic circRNAs (tricRNAs), which are derived from the linear intron of tRNA precursors.⁷Fourth, small nuclear RNA (snRNA) can be transcribed from a promotor and encoded within intronic sequences; its vital function is in splicing of introns from primary genomic transcripts.⁸Fifth, small nucleolar RNAs (snoRNAs), encoded in the introns of protein-coding or non-coding genes and co-transcriptionally induced with host genes, regulate mRNA processing and are involved in rRNA processing, stress response, and metabolic changes.⁹Sixth, small interfering RNA (siRNA) is involved in the simultaneous synthesis of mature siRNA with transcription by RNA POL II, RNA POL III, or RNA POL IV, creating dsRNA, temporarily silences desired genes, and triggers RNAi upon binding to the target transcript.¹⁰Seventh, PIWI-interacting RNA (piRNA or piwiRNA) is transcribed inside the nucleus and exported to the cytoplasm, where it is further processed to form mature piRNAs that get loaded onto PIWI proteins, which have crucial roles in silencing transposons, preserving germline DNA integrity and epigenetic regulation of sex determination, and generating heterochromatin.¹¹ Non-coding RNAs can cross-talk and interact with coding genes, proteins, and other ncRNAs (LncRNA and ceRNA),¹² and when they are simply transcriptional noise it means that their expression levels remain nearly the same when observed among different tissues.¹³ However, many ncRNAs are also detected in the cytoplasm and nucleus. They interact with mRNAs, genes, and other ncRNAs, play regulatory functions, and show integrated regulatory networks with the ceRNA network and ncRNA-mediated networks in genomics and proteomics.¹⁴ This Review aims to highlight the significant role of ncRNAs, decipher the different types of RNAs and their role as biomarkers in diseases, and focus on the bioinformatics tools and the available databases covering the wide utility of different types of ncRNA databases as well as the disease- and cancer-associated databases.

Regulation of Non-coding RNAs

Regulation of non-coding RNAs occurs at multiple levels by its interactions with DNA, RNA, and proteins, as illustrated in Figure 2. MicroRNA interacts with the 3′UTR of target mRNAs, 5′UTR, coding sequence, and gene promoters. There are various mechanisms underlying miRNA-mediated gene regulations, including gene silencing via the miRISC complex, transcriptional and post-transcriptional gene regulation within the nucleus, and translation activation. Studies revealed that the dynamical behavior of microRNA contributes gene regulation, including functionalized compartmentalization and shuttling of miRISC within the cells.¹⁵LncRNA modulates in several ways, such as chromatin regulation, where protein-lncRNA localization and functions are mediated by recruitment and decoy of chromatin modifiers and direct interaction of proteins-lncRNA in cis and trans positions on chromatin; transcriptional regulation, with gene silencing by lncRNA, where lncRNAs transcribe at enhancers, regulatory networks involving cis-acting lncRNAs, chromatin scaffolding, and nuclear condensation; and post-transcription regulation, via RNA splicing, stability, and translation, cellular mechanisms where lncRNAs are localized to specific organelles, such as exosomes and mitochondria, and cellular homeostasis is dependent on the action of lncRNAs.¹⁶Small nucleolar RNA plays roles in rRNA, tRNA, and mRNA modifications. It guides N4-acetylcytidine, 2′-O-methylation, and pseudouridylation of rRNAs and regulates alternative splicing and levels of mRNA like a miRNA.¹⁷Small interfering RNA (cleavage product of dsRNA) mediates gene silencing (specific degradation of mRNA), shown to be post-transcriptional sequence-specific process.¹⁸Circular RNA modulates gene expression in the nucleus, which then act as microRNA decoys, regulate transcription, splicing, and chromatin interactions, can sequester proteins, and function as protein scaffolds.¹⁹PiwiRNA regulates transposon silencing and pachytene piRNA function and can distinguish self from non-self and fight viral infections.²⁰

Figure 2.

Depicting the types of regulation by different ncRNAs (microRNA, lncRNA, circRNA, piRNA, siRNA, snoRNA) inside the cell. Dashed line arrow shows the regulation by the respective ncRNA, and dark line arrow shows the normal processes going on inside the cell.

Non-coding RNA Databases

Typically, data exists in a raw format characterized by diverse structures, necessitating a preprocessing stage to render the data suitable for algorithms, tools, and tailoring to the specific context of the relevant field. Techniques for data integration are employed to manage disparate resources, facilitating the consolidation of all data into organized databases. Tools and pipelines used for data integration include dChip-GemiNi (gene and microRNA network-based integration), MAGIA2, mirConnX, and IntegraMiR. Transcriptomic databases store information related to transcripts of the human genome whether coding mRNAs or ncRNAs, genes, and transcription factors.²¹

Training, testing, and verifying the input data for unique prediction are done by using machine learning (ML) approaches such as SVM, Monte Carlo algorithm, etc. Some NLPs (Natural Language Processing) help assess biological data. Biological data formats represent biological information in a file. A set of commonly used formats (FASTA/Q, SAM, VCF, GFF/GTF, etc.) are used for importing and exporting data.²²

Databases

According to the 2018 NAR Molecular Biology Database Collection report, the current number of databases is 1737. However, this comprehensive list covers the broad spectrum of different primary and secondary biological databases containing multiple types of “omics” data.²³ We emphasize four primary approaches in our database search: (1) acquiring information about databases, (2) making predictions through the integration of data within databases, (3) employing deep learning methods for the analysis of ncRNA, and (4) making identifications through the analysis of genomic data that reports on ncRNA, mRNA, and/or protein expression. Multiple platforms (databases and tools) are utilized for a complete systems-level analysis workflow, which causes hurdles and produces a fragmented user experience. Having many objectives, users analyze data on different platforms, which requires users to learn the details of each interface successfully, and they must merge the data to get useful results. When the users are inexperienced, they pose barriers while doing analysis. Moreover, significant data may be omitted during the analysis, and this is due to the disintegration of sources and analysis platforms. To facilitate multi-platform data analysis and interpretation, developers designed databases/tools to be more user-friendly and to increase the user experience.²⁴

Non-coding RNAs play a key role during the transcriptional and post-translational processes. It is challenging to perform experimental processes to understand diversity in the functions of ncRNA. Computational approaches performed to analyze ncRNAs based on in silico approaches include annotation, data mining of new RNA datasets, and databases and tool resources.²⁵ Computational methods are developed to ensure that valid ncRNA candidates should be found with a high confidence score, precision, and accurate significant value using statistical models for further experimental validation. ncRNA is currently a hot topic for studying drug targets and checking their correlations with diseases. Several reviews have already been published on computational databases and tools. However, many of them are outdated, with the rise of ncRNAs, and updating the list of computational databases and tools for ncRNAs is required from time to time. Computational predictions offer an alternative source of insight into ncRNA function, complementing experimental findings. The knowledge acquired from experiments continuously informs and refines computational methods, leading to more accurate and insightful predictions. In turn, these predictions shed light on potential interactions between ncRNAs and other biological components, guiding further experimental validation. This synergistic relationship between computational and experimental approaches fosters a deeper understanding of ncRNAs and their roles in various biological processes. The emphasis is on offering an updated database and tools, sparing researchers the challenge of navigating vast repositories. Current transcriptomics methods and computational resources in the ncRNA field enhance the classification, annotation, identification, storage, prediction, and analysis of ncRNAs.²⁶

Furthermore, ncRNA databases are public repositories utilized by researchers. Some of the commonly used genomic databases, such as TCGA, ICGC, PCAWG, Cancer Cell Line Encyclopedia (CCLE), Gene Expression Omnibus (GEO), NCI Genomic Data Commons (GDC), cBioPortal, UALCAN, KM plotter, NCBI, ENCORE, Refseq, UCSC, PubMed, PDB, GENCODE, COSMIC, Ensembl, ICGC, RNAcentral, etc., work as the source of ncRNA databases and tools. Figure 3 illustrates that deep learning, new networks, and statistical analysis can form new ncRNA databases from source databases.²⁷

Figure 3.

This illustration shows that data is extracted from source databases. Deep learning, statistical models, and network studies are done and ncRNA databases are curated. Note: In this figure, only a few representative source databases and ncRNA databases are shown.

We classify databases and tools into several categories: basic genomic annotation, expression profile, molecular interactions/regulatory networks, identification, sequence variants, prediction, and disease association, as tabulated in Table 1. In the table, enlisted databases and tools have various functions from decoding the functions and identification to prediction; categorization helps in studying various aspects, such as the molecular interactions of ncRNA-DNA, ncRNA-RNA, ncRNA-protein, LncRNA-miRNA, miRNA-mRNA, ceRNA networks, sequence variants of single nucleotide polymorphism (SNPs), SNPs in structure, and SNPs in interactions; expression profiles of expression patterns, epigenetics, secondary structure, TFBS (transcription-factor binding site) in ncRNAs, mRNA coexpression; basic genomic annotation of biological function, types and location, evolutionary conservation, sequence information, coding potential; disease association and its role in progression, and diagnostic and prognostic markers for building therapeutic targeting strategies. This tabular data can assist researchers in getting a better understanding of the ncRNA databases. Some of the databases and tools are for more than one type of ncRNAs; those specific to one type of ncRNA are called Multiple-Class RNA Databases. Some of the databases are mostly used by researchers as they are more user-friendly and more convenient. The most significant cutting-edge public repositories of ncRNA data are as follows: for Multiple RNA class, Rfam, RNACentral, NONCODE, deepBase, ncRNAdb, NPInter, NSDNA; for microRNA, miRBase and miRTarBase; for lncRNA, NONCODE, LncRNAWiki, LncBook, and LNCipedia; for circRNA, circBase, CircAtlas, and CircInteractome; and for piwiRNA, piRbase and piRNAdb.²⁸⁻³⁰ With the increasing prevalence of in silico studies, researchers can leverage computational sources, such as databases, for rigorous investigations. This Review delineates recent advancements in databases, bioinformatic tools, and innovative in silico approaches that facilitate the prediction and establishment of biological interactions involving ncRNAs. Special attention is given to ncRNA species, particularly those identified in humans.

Table 1. Databases Categorized into Basic Annotation, Expression Profile, Molecular Interactions/Regulatory Networks, Identification, Sequence Variants, Prediction, and Disease Association^a.

^aColor representation for different ncRNAs is different.

Databases and Tools for Predicting Interactions

Databases and tools that predict the interactions between RNA and RNA, or RNA and proteins, and any other interactions are discussed as indispensable resources for understanding the complex regulatory networks within cells. RAID is a resource that integrates experimental and computational prediction of RNA-associated (RNA-protein/RNA-RNA) interactions involving various RNAs (including circRNA, lncRNA, miRNA, mRNA, miscRNA, pseudogenes, rRNA, scRNA, sncRNA, snoRNA, snRNA, sRNA and tRNA) and contains many species. RNAInter means “RNA interactome” database and promotes research on RNA interactions with RNA, DNA, protein, and compounds in disease processes. Annotation in RNAInter is shown with disease associations and tissue-specific expression. ViRBase investigates viral infections via means of viral and host ncRNA-associated interactions. NPInter is newly updated with all recently identified ncRNA interactions, which were curated manually, and each interaction entry shows detailed annotations and prediction scores. ChIRP-seq data is captured for investigating the circRNA interactions and ncRNA-DNA interactions. The new interface is more convenient than StarBase and RAID, and to add confidence to the interactions, the predictive scores and visualization modules are integrated. SomamiR shows somatic mutations in miRNA and their target sites, as well as their interaction with competing endogenous RNAs (ceRNAs) like mRNA, lncRNA, and circRNA. A list of somatic mutations in miRNA seed regions gives information about miRNA target recognition. The miR2GO web server is integrated with SomamiR 2.0 for finding the functional impact of these mutations in miRNA target regions, and other sources such as GWAS and the biological pathway are added for functional analysis of somatic mutations altering miRNA-ceRNA interactions. ENCORI (The Encyclopedia of RNA Interactomes), previously known as StarBase, is a well-known database that integrates different data for mainly RNA species from high-throughput sequencing studies. ENCORI focuses exclusively on miRNA-ncRNA, miRNA-mRNA, RBP-ncRNA, and RBP-mRNA interactome data for visualization, while also incorporating complementary studies such as those involving survival and differential expression analyses. The miRFunction and ceRNAFunction web servers were developed to predict the function of miRNAs and other ncRNAs from the miRNA-mediated regulatory networks. LncBook adds three tools, RNAhybrid, miRanda, and TargetScan, to predict more lncRNA and miRNA interactions. LncFinder was released as a web server and R package that helps in finding the properties of lncRNAs and mRNAs, predicting lncRNA-protein interactions, and analyzing lncRNA evolution. Circ2Traits (circular 2 traits) is a database that focuses on two traits: (1) identification of the association between circRNAs and miRNAs specifically related to diseases in humans so that how circRNA nteract with disease and enrichment genes can be calculated; (2) SNPs associated with human diseases and their association with circRNAs. piRNAclusterDB shows the interactions among PIWI-interacting RNAs (piRNAs) and PIWI proteins. A high divergence of genomic piRNA-producing loci has been seen, along with their regulation and evolution of piRNA-producing loci across various tissues and species.

Non-coding RNAs as Therapeutic Biomarkers

Nowadays ncRNAs can be manufactured as synthetic oligonucleotides, which can bind to an mRNA and inhibit it directly and thus can modulate the activity of targets.³¹ Non-coding RNAs play a wide range of roles in the onset and progression of cancer. Dysregulated ncRNAs serve as hallmarks of cancers as oncogenes or tumor-suppressor genes and hence can be used as newer strategies to uncover the role of “dark matter” in the genome in cancer.³² They represent potential diagnostic and therapeutic biomarkers in hereditary diseases as well, like in alpha thalassemia, Alzheimer’s disease, Angelman syndrome, Beckwith–Wiedemann syndrome, metaphyseal chondrodysplasia, facioscapulohumeral muscular dystrophy (FSHD), Hirschsprung disease (HSCR), Huntington’s disease, Opitz–Kaveggia syndrome, Prader–Willi syndrome, pseudohypoparathyroidism type 1b (PHP1b), Silver–Russell syndrome, spinocerebral ataxia type 7 (SCA7), spinocerebral ataxia type 8, spinal muscular atrophy, heart diseases (myocardial infarction, hypertension, cardiac developmental disorder, cardiac fibrosis), cancer (breast cancer, cervical cancer, lung cancer, leukemia), and viral infections (Dengue virus, hepatitis C virus, Japanese encephalitis virus, enterovirus 71 (EV71), human immunodeficiency virus, SARS-COV,³³ cerebral ischemia).³⁴ Nemeth et al. provide an overview of the functions of miRNAs, lncRNAs, and circRNAs in cancer and various other significant human diseases, including cardiovascular, neurological, and infectious diseases. They also discuss the potential applications of ncRNAs as disease biomarkers and therapeutic targets.³⁵

Role of Non-coding RNAs in Cancers and Various Diseases

Non-coding RNAs play a role in many physiological processes, from regulating healthy cells to developing cancerous cells. Cancer develops gradually, characterized by uncontrolled proliferation of cancer cells, altered apoptosis mechanisms, and infiltration of normal cells, leading to metastasis, a primary cause of death. Following metastasis, treatment becomes more challenging due to increased risk factors. Non-coding RNAs (ncRNAs) are implicated as both oncogenic factors and tumor suppressors across various cancers. Their roles are often multifaceted, with interactions among different ncRNAs regulating crucial cellular processes, including those involved in cancer, influencing its initiation and progression through three mechanisms, acting as oncogenes or tumor suppressors or impacting metastatic processes.³⁶

Additionally, the distinctive tissue- and cancer-specific expression patterns of ncRNAs suggest their potential utility as diagnostic and prognostic biomarkers. However, a deeper understanding of ncRNAs, including newly identified tsRNA and circRNA, through comprehensive characterization and functional analysis is necessary to elucidate their molecular mechanisms in tumorigenesis. This exploration could unveil promising therapeutic targets for cancer treatment.³⁷

“Jumping genes” or transposable elements (TEs) are considered essential factors for disturbing several molecular processes that cause the plasticity and evolution of the genome. Alu and LINE-1 repeats are prevalent in humans and may have their molecular roots in numerous clinical disorders.³⁸ Transposons play a role in cancers by different mechanisms: 1. gene expression alteration by insertion; 2. alternative splicing; 3. double-strand breaks; 4. onco-exaptation; 5. placental gene expression; 6. expression of oncogenic LncRNAs; and 7. transposons in autophagy and senescence. Elucidating the expression of oncogenic LncRNAs reveals the influence of TEs on ncRNAs. TEs are present as part of lncRNAs, providing sites for transcription initiation and post- transcriptional modifications. Long terminal repeat (LTR) retroelements are part of the lncRNA example HOST 2 (Human Ovarian cancer Specific Transcript 2), which is overexpressed in triple-negative breast cancer (TNBC) and pancreatic cancer. LTR12C acts as a promoter for SchLAP1 lncRNA and overexpresses SchLAP1 in prostate cancer. Some of the LTRS promote oncogenic lncRNAs, including BACE1 and HULC in liver and breast cancer. Continuous research on jumping genes can give us a better understanding of their biological roles and applications in cancers.³⁹ Beňačka et al. gave examples of ncRNAs involved selectively in some cancer diagnostic indications such as breast cancer, brain tumors, papillary thyroid carcinoma, hepatocellular carcinoma, intrahepatic cholangiocarcinoma, bronchogenic carcinoma, kidney cancers, bladder carcinoma, prostate cancer, cervix/endometrium/ovarian cancer, gastric cancer, colorectal cancer, and leukemia and non-cancer diagnostic indications such as gastritis, Crohn’s disease, ulcerative colitis, brain/spinal injuries, Alzhemier’s disease, Parkinson’s disease, diabetes mellitus, coronary artery disease, heart failure, COPD, asthma, NAFLD, cirrhosis, and liver failure.⁴⁰

Non-coding RNAs transferred within the body in bodily fluids (serum, plasma, urine, saliva, and others), called extracellular (EC) or circulating ncRNAs (EC ncRNAs), mediate extracellular communication; they can also be called non-exosomal ncRNAs. Other EC ncRNAs are transported inside the cells for intercellular communication, enclosed by extracellular vesicles (such as exosomes, microvesicles, and apoptotic bodies); they can also be called exosomal ncRNAs. Exosomal and non-exosomal ncRNAs that regulate physiological homeostasis and pathological events in health and disease have been examined.⁴¹ NcRNAs can form integrated networks in major diseases, thereby working in chromatin modulation, splicing regulation, crosstalk with other ncRNAs in proteomics, and antisense transcription.³³

Diseases and Cancer-Associated Non-coding RNA Databases

The versatile roles of ncRNAs span a broad spectrum of cellular processes, with some exerting significant influence in disease contexts. Despite substantial strides in recent decades, the identification of numerous potential ncRNA targets remains incomplete. The experimental validation of all prospective targets is a laborious and costly process, prompting increased interest in alternative approaches, particularly those leveraging bioinformatics. Given the involvement of ncRNAs in biological functions and signaling pathways, their dysregulation has been linked to various human diseases, including cancer. Recognizing the pivotal role of ncRNAs in disease contexts is crucial for biomedical researchers seeking diagnostic, therapeutic, or preventive solutions for specific disorders. This underscores the clinical relevance of understanding ncRNAs. Notably, dedicated databases have emerged to consolidate information on ncRNAs’ involvement in various diseaprovidses, providing a valuable resource for researchers (Table 1).

There are various ncRNA databases associated with diseases. For instance, Lnc2Cancer is an interactive analysis platform including single cell, RNA sequencing expression data, and associations between lncRNA or circRNA and various human cancers. LnCAR (LncRNAs from cancer arrays) overcomes the derivation of reannotation of public microarray data by providing expression profiles and prognostic landscape of lncRNAs. Circad (circRNAs associated with diseases) deals with the association of circular ncRNAs with diseases with reference to International Statistical Classification of Diseases (ICD) codes, providing standard disease nomenclature. CircRiC (circular RNA in cancers) provides an analysis of circRNAs and prediction for their being RNA regulatory elements as biogenesis regulators on drug response and association with mRNA, proteins, and mutations. MiOncoCirc helps in finding onogenic circRNAs. The iPiDA-GCN predictor discriminates features from similarity networks of piRNA and disease association patterns and reveals the potential pathogenesis at the RNA level affected by piRNA-disease associations. GeneCaRNA is an all-inclusive compendium of ncRNA genes that empowers ncRNA interaction networks related to diseases as well. ncRPheno databases identify and validate diseases related to ncRNAs, and miRdisNET discovers microRNA biomarkers and predicts potential microRNA-disease associations. cancerEnD is available to identify new enhancer biomarkers for therapy and disease diagnosis. The MetaGxData compendium integrates three packages containing curated and processed expression datasets for breast (MetaGxBreast), ovarian (MetaGxOvarian), and pancreatic (MetaGxPancreas) cancers. ncRNAVar association data between validated ncRNA variants and human diseases. MNDR (Mammalian ncRNA-Disease Repository) integrates manually curated data from literature that experimentally supports and predicts ncRNA-disease associations. The SPENCER database investigates ncPEPs (small peptides encoded by ncRNAs) in multiple cancers, OMCD compiles data for microRNA gene expression of normal and tumor tissues, and the BioXpress database integrates RNA-seq-derived gene expression for pan-cancer analysis. ViRBase investigates viral infections via means of viral and host ncRNA-associated interactions.

Some are site-/organ-specific databases. OncoDB is for analysis of gene expression and viral infection in cancer. The lung cancer database, LCE, explores gene expression and clinical associations. The breast cancer databases include BreastMark, which is a powerful tool for examining putative gene/microRNA prognostic markers; bc-GenExMiner for breast cancer differential gene expression analyses; BC-TFdb, for transcription factors that regulate breast cancer; and BCLncRDB for lncRNAs expression. The liver cancer database CancerLivER maintains gene expression datasets of liver cancer along with the putative biomarkers. The type 2 diabetes database T2DB is for long ncRNA genes expression in Type II diabetes. FibroDB identifies lncRNA gene expression during fibrosis. Three databases were found that provide ncRNA associations with the immune system: the RNA2Immune database identifies ncRNA-immune system associations, ImmunemiR prioritizes immune-microRNA disease associations, and ncRI provides experimentally validated ncRNAs in inflammation. These disease- and cancer-associated databases are depict in Figure 4.

Figure 4.

Depiction of the disease- and cancer-associated ncRNA databases.

Advantages and Major Challenges of Databases

The availability of databases offers researchers a continuous influx of information through regular updates and data derived from meticulous experimental processes, fostering the design of experimental trials and the exploration of new interactions. Computational and statistical methods enable predictions, with various databases and tools, such as RNAInter, oRNAment, NPInter v4.0, miRDB, and ENCORI, incorporating in silico prediction algorithms. The integration of deep learning, mimicking human brain function through artificial neural networks, has significantly impacted omics studies, introducing advancements like deep structured learning and hierarchical learning for analyzing quantitative data properties. The era of next-generation sequencing has provided unprecedented opportunities for identifying new ncRNAs. Cheng et al. discuss the application of bioinformatics in studying ncRNAs and highlight major challenges in computational analysis.^42,43

To enhance the impact of computational models in biology, addressing challenges such as comprehensiveness, accessibility, reusability, interoperability, and reproducibility is crucial. Computational models play a role in studying the dynamic behavior of complex systems, and improvements involve proper model annotations, community-supported standardization of formats, interoperability through implementing standards in tools and platforms, transparency via comprehensive documentation across modeling frameworks, and the use of scorecards and simple rules to enhance model reproducibility.⁴⁴

Conclusion

The field of ncRNA research demonstrates the diversity of nature and its complexity in cellular biology. Advancements in bioinformatics, sequencing technologies, proteomics, and microarrays have identified various ncRNAs. Understanding the unique functions of these ncRNAs is challenging due to their complex networks and biological pathways. The use of ncRNA therapies in drug development will increase, and further information is needed to understand their pharmacokinetics and dynamics. This Review covers various databases and tools and helps researchers by reducing the extensive research needed for selecting appropriate databases and tools. Non-coding RNAs have recently become strong therapeutic prospects in light of recent discoveries that these ncRNAs have an extensive function in disease progression and general pathophysiology. These databases will be invaluable resources for researchers working in this field. There are still unresolved issues concerning ncRNA research, even with the abundance of resources available. Future research should focus on ML models that recognize disease patterns. The use of ncRNAs may increase in the coming years, contributing to successful precision medicine and personalized therapies.

Glossary

Abbreviations

ncRNA

Non-coding RNA

piRNA

PIWI-interacting RNA

lncRNA

Long non-coding RNA

circRNA

Circular RNA

miRNA

MicroRNA

SVM

Support vector machine

RNA POL

RNA polymerase

TCGA

The Cancer Genome Atlas Program

ICGC

International Cancer Genome Consortium

GENCODE

Genome research and part of the ENCODE

COSMIC

Catalogue of Somatic Mutations in Cancer

ENCORE

ECHO Native Circadian Ontological Rhythmicity Explorer

UCSC

University of California Santa Cruz

PDB

Protein Data Bank

CSCD

Cancer-Specific CircRNA Database

ceRNA

Competing endogenous RNA

CCLE

Cancer Cell Line Encyclopedia

The authors declare no competing financial interest.

Special Issue

Published as part of ACS Pharmacology & Translational Sciencevirtual special issue “Nucleosides, Nucleotides, and Nucleic Acids as Therapeutics”.