miRiadne: a web tool for consistent integration of miRNA nomenclature

Abstract

The miRBase is the official miRNA repository which keeps the annotation updated on newly discovered miRNAs: it is also used as a reference for the design of miRNA profiling platforms. Nomenclature ambiguities generated by loosely updated platforms and design errors lead to incompatibilities among platforms, even from the same vendor. Published miRNA lists are thus generated with different profiling platforms that refer to diverse and not updated annotations. This greatly compromises searches, comparisons and analyses that rely on miRNA names only without taking into account the mature sequences, which is particularly critic when such analyses are carried over automatically. In this paper we introduce miRiadne, a web tool to harmonize miRNA nomenclature, which takes into account the original miRBase versions from 10 up to 21, and annotations of 40 common profiling platforms from nine brands that we manually curated. miRiadne uses the miRNA mature sequence to link miRBase versions and/or platforms to prevent nomenclature ambiguities. miRiadne was designed to simplify and support biologists and bioinformaticians in re-annotating their own miRNA lists and/or data sets. As Ariadne helped Theseus in escaping the mythological maze, miRiadne will help the miRNA researcher in escaping the nomenclature maze. miRiadne is freely accessible from the URL http://www.miriadne.org.

Figure 1.

miRiadne main features. The five miRiadne's features grouped in two main classes: the first class includes the four Rosetta features used to translate miRNA names (Rosetta Stone functions) and data sets (Rosetta Data functions) produced by miRNAs detection platforms: they are the Translation, Overlap, Intersection and Update functions. The second main feature is the Time Warp that allows one to browse any miRNA name or name lists evolution across miRBase versions, in any species.

Figure 2.

Name-driven miRNA list comparison. (a) Overlap of miRNA signatures from two different papers (18,19). Overlap between miRNA names as they are reported in the papers (Venn diagram on the left) shows 16 commonly detected miRNAs, while overlap between miRNA signatures with updated annotations (Venn diagram on the right) shows 15 common miRNAs. Further overlap between the two commonly detected miRNAs uncovers that only 11 miRNAs are not affected by inconsistent annotations between the two papers and are correctly included in the comparison in both cases, and one miRNA (hsa-miR-923, in red) is not present in the updated list. (b) The 15 miRNAs commonly detected by the two works in the result table obtained with the ‘Rosetta Intersect’ function. Names are automatically updated to latest miRBase version, mature sequences appended and retired miRNAs (such as hsa-miR-923) highlighted and excluded. Destiny of retired miRNAs can be further investigated with the Time Warp function.