(Georgia Institute of Technology, 2022-05)
McCann, Holly
Non-coding RNAs play key roles in cellular systems such as transcription and translation. Computational predictive methods provide insight into the secondary structures of non-coding RNAs from diverse organisms for which there are no high-quality crystal structures. This research project builds upon the existing framework of the R2DT (RNA 2D Templates) tool developed by the RNAcentral consortium with collaboration from Georgia Tech researchers. R2DT predicts RNA secondary structures and displays them in the form of a 2D topology diagram by using known, related structures as templates for prediction and visualization. The automated template-based approach, however, often ignores unique species-specific regions such as ribosomal RNA expansion segments.
To improve the functionality of R2DT, the secondary structures of insertion regions which are not found in any template structure are now predicted using the RNAfold algorithm developed by ViennaRNA. This hybrid approach can generate more accurate structures for various types of RNA molecules with unique features across entire phylogeny. In addition to improvements to the folding algorithm, a new front-end for R2DT has also been developed which builds on the PDBe RNA Viewer. The generated topology diagrams now display RNA secondary structures represented either as a contour line or as individual nucleotides. The improved RNA Viewer provides support for visualization of both canonical and non-canonical base pairs (using the symbolism proposed by Leontis and Westhof), and includes interactive features such as object highlighting and a tooltip. These developments improve the accuracy and functionality of computationally predicted RNA topology diagrams. The applet further enables users to visualize various structural and evolutionary data and to easily generate publication-quality images of data mapped onto secondary structures. This applet has been integrated with the PDB MSA and Mol* viewers to display non-coding RNA molecules and associated data in three dimensions on the RiboVision2 webserver.