RNABP COGEST: a resource for investigating functional RNAs

Abstract

Structural bioinformatics of RNA has evolved mainly in response to the rapidly accumulating evidence that non-(protein)-coding RNAs (ncRNAs) play critical roles in gene regulation and development. The structures and functions of most ncRNAs are however still unknown. Most of the available RNA structural databases rely heavily on known 3D structures, and contextually correlate base pairing geometry with actual 3D RNA structures. None of the databases provide any direct information about stabilization energies. However, the intrinsic interaction energies of constituent base pairs can provide significant insights into their roles in the overall dynamics of RNA motifs and structures. Quantum mechanical (QM) computations provide the only approach toward their accurate quantification and characterization. 'RNA Base Pair Count, Geometry and Stability' (http://bioinf.iiit.ac.in/RNABPCOGEST) brings together information, extracted from literature data, regarding occurrence frequency, experimental and quantum chemically optimized geometries, and computed interaction energies, for non-canonical base pairs observed in a non-redundant dataset of functional RNA structures. The database is designed to enable the QM community, on the one hand, to identify appropriate biologically relevant model systems and also enable the biology community to easily sift through diverse computational results to gain theoretical insights which could promote hypothesis driven biological research.

Figure 1.

(A) Matrix table for efficient browsing of base pair information available in RNABP COGEST. Each cell of the matrix indicates a specific base–base interaction in a particular geometry, e.g. the cell highlighted by the blue box indicates W:W type interaction between two Adenine bases. Selecting the radio button of a cell, followed by clicking on the search button at the end of the table, opens up a list of all possible base pairs between the given bases in their selected geometries. Cells without radio buttons (example cells marked in red) correspond to base pairs and geometries which cannot be stabilized through hydrogen bonds. (B) Pie chart showing distribution of occurrence frequencies of seven ‘most frequent’ base pairs in the non-redundant dataset. (C) Distribution of occurrence frequencies of base pairs, referred to as ‘others’ in (B), and, which occur more than 100 times in the non-redundant dataset.

Figure 2.

Intrinsic interaction energy of A-I/CG triad (top right) is higher than that of A-I/GC triad (top left), though the latter shows greater stability during simulation. This may be explained (cf. text) in terms of the variation in contribution of the correlation term (dispersion) for G:A S:S Trans base pair in the A-I/GC triad, and that for the C:A S:S Trans base pair in the A-I/CG triad. Images and table are adapted from Réblová et al. (52).