miRNAs are a class of short (~22 nucleotide) noncoding RNAs expressed in both animal and plant cells. These small RNA molecules bind preferably to the 3’ untranslated region (3’UTR) of protein coding genes and make them degradation and/or translation inhibition
]. A typical miRNA target site has perfect or nearly perfect complementarity to the miRNA so-called seed sequence, i.e., ~7 nucleotides from miRNA 5’ end
]. Both experimental and computational studies have shown that miRNAs are a class of post-transcriptional regulators, and play important roles in cellular process
], developmental control
], hormone secretion
], cell differentiation and proliferation
], and responses to environmental stresses
]. One challenge facing miRNA research is to accurately identify the target mRNAs, because of the very limited sequence complementarity between miRNAs and their targets, and the scarcity of experimentally validated gene targets to guide accurate prediction models
, a small flowering plant with only several weeks of life cycle, has important advantages for the research of genetics and molecular biology, thus some special databases for Arabidopsis thaliana
have been established. TAIR is a comprehensive information system that deposits genome, expression profiling, proteome, pathway and functional annotations, together with many analysis and visualization tools
]. For miRNAs, Sanger Institute’s miRBase serves as the central depository where miRNAs are experimentally validated. The current release, version 10.0, contains over 266 Arabidopsis thaliana
]. However, the functions of most miRNAs are still unknown so far. The Arabidopsis Small RNA Project Database, abbreviated as ASRP, intends for Arabidopsis thaliana
small RNA mature sequences, transcripts and locus
]. PMRD (Plant miRNA Database) has collected miRNA sequences, miRNA target genes and expression profiles of some model plants
] collects information about the expression profiles of 190 Arabidopsis thaliana
pri-miRNAs in seven different developmental stages and provides simultaneous comparison of expression levels between various microRNA genes in diverse organs and developmental stages.
However, all databases mentioned above pay main attention to collecting miRNA-related information such as miRNA sequences, miRNA-target bindings. No system has been devoted to the functional exploration of miRNAs. Even though a few databases include expression information of miRNAs, the coverage is quite limited, failing to integrate most of the microarray and high-throughput experimental results. The recently published database miRNEST
] is a comprehensive collection of animal, plant and virus microRNA-related data, including miRNA sequence, expression, polymorphisms and targets data, but it has no function annotations and pathways data, and does not provide comparative analysis of expression profiles from identical tissues or samples between miRNAs and their targets.
In this paper, we introduce miRFANs, an integrated database with a friendly web interface for functional annotations of miRNAs, which encompasses miRNA-target interactions, expression, functional annotations, pathway terms. For miRNA targets, three predicted sets by three computational algorithms (psRNATarget
] and UEA target-finder
]) and two experimentally validated sets (miRTarBase
] and TarBase
]) are collected. However, the miRNA-target interactions predicted by computational methods only imply the possibility that miRNAs regulate their targets under certain spatio-temporal conditions, as the regulation of miRNAs is well known to be condition- and tissue-specific. The expression data provides us the chance to evaluate the effect of miRNA binding in the sense of a certain environmental condition and tissue, on the basis of several evidences showing that the expression profiles of miRNAs are closely related to that of their target genes. The reciprocal expression patterns between miRNAs and their targets can be remarkable evidence of miRNA regulatory events. Therefore, we download expression datasets of both miRNAs and genome from the gene expression omnibus (GEO)
]. Expression datasets are preprocessed and integrated for readily identifying co-expressed or differentially expressed mRNAs and their targets. We also developed a web interface supporting diverse query entries that include search by miRNA, expression profile or pathway, and a mining toolbox including correlation, differential expression analysis and clustering to explore the expression data.
Thus, miRFANs can serve as a comprehensive resource for exploring the functions of Arabidopsis thaliana miRNAs.