ابزارهای بیوانفورماتیک در تحقیقات IncRNA Bioinformatics tools for lncRNA research

۱٫ Introduction Recent high throughput sequencing technologies have enabled us to obtain a number of candidates of long non-coding RNAs (lncRNAs). However, because the current experimental identification methods are still limited in their throughput, fast bioinformatics tools to identify and characterize lncRNAs with reasonable accuracy are required. Because non-coding RNA is an exclusive category of RNAs that do not code for functional polypeptides, the first task for bioinformatics is to identify lncRNAs by screening long transcripts that do not seem to code for proteins. The objective of lncRNA research, however, is not only to find long non-coding RNAs but also to identify their functions. There are various bioinformatics tools for predicting the structures and functions of RNA sequences, including several tools that incorporate other experimental data in the analysis, but it is not obvious which tools are most useful for any particular objective. It is known that structures, especially secondary structures, are important determinants of the functions of non-coding RNAs. It is also observed that genomic elements sharing similar functions are conserved between species. Therefore, secondary structures and their conservation are examined using bioinformatics tools to try to determine their functional categories. The predictions of secondary structures, however, are not always accurate. Nevertheless, although it is not always easy to extract concrete structural motifs related to functions, functional domains still may have structural features. Important clues for the functions of lncRNAs, including when, where and with what they are used, can be extracted from experimental data. Spatiotemporal expression patterns (in tissues, subcellular compartments, and differentiation/developmental stages) by RNA-seq or microarray indicate when and where functions are activated. Co-expression analysis with protein coding genes is useful for predicting with what, but a more direct way is to detect the interactions with proteins and other RNAs. Interactions with proteins may indicate the type of the function; furthermore, complementary bases in two RNA molecules often form base-pairs, giving high sequence specificity for the target RNAs of the functional RNAs. RNA–RNA interactions can be screened by searching reverse complementary subsequences, but precise analysis of structures both within and between RNA molecules is necessary for accurate prediction. In this paper, we review available bioinformatics tools for research into lncRNAs, including their discovery, analyses and predictions of the secondary structures, conservation, interactions with other RNAs and proteins, co-expression with protein-coding genes, tissue-specificities, and subcellular localizations. We also consider useful databases. ۲٫ Finding long non-coding RNAs There are two steps in the identification of lncRNAs. In the first step, the transcribed units of the lncRNAs are identified. The fragments of the transcribed RNA sequences, observed by using next-generation sequencing (NGS) technologies or tiling microarrays, are mapped to the reference genome and summarized to obtain the transcribed units of the RNAs. The second step classifies the transcribed units as coding or non-coding: the sequences of transcribed units are evaluated on the basis of codon statistics and similarity to known protein sequences. Before NGS technologies became available, however, it was common to predict candidates of (functional) non-coding RNAs on the basis of their sequences and to experimentally verify their expression. For this prediction, conserved features (including secondary structures) of candidate sequences are considered. These analyses are still important for characterization of the functions of lncRNAs.