CS-Annotate: A Tool for Using NMR Chemical Shifts to Annotate RNA Structure

Abstract

Here, we introduce CS-Annotate, a tool that uses assigned NMR chemical shifts to annotate structural features in RNA. At its core, CS-Annotate is a deployment of a multitask deep learning model that simultaneously classifies the solvent exposure, base-stacking and -pairing status, and conformation of individual RNA residues from their chemical shift fingerprint. Here, we briefly describe how we trained and tested the classifier and demonstrate its application to a model RNA system. CS-Annotate can be accessed via the SMALTR (Structure-based MAchine Learning Tools for RNA) Science Gateway (smaltr.org).

1H, 13C and 15N assignment of stem-loop SL1 from the 5'-UTR of SARS-CoV-2

“…Qualitative evaluation of glycosidic torsion angles via the intensity of the intra-base H1′–H6/H8 NOESY cross peak did not reveal a tendency for syn conformation for any of the internal loop nucleotides. Furthermore, global chemical shift analysis using CS-Annotate (Zhang et al 2021 ) supported a largely stacked arrangement of all nucleobases of the internal loop, except for C28 (SI Fig. S2).…”

NMR of RNA - Structure and interactions

“…Several approaches with a combination of structure prediction programs, MD and NOEs, or chemical shift information were used to produce structural ensembles of HIV-1 TAR [67] or SINEB2 RNAs [68]. Chemical shifts were suggested to be used to refine RNA structures [69] or decode structural features, such as solvent exposure, base stacking, or base pairing [70]. Liu et al have identified, designed, and assigned sequences with no or limited chemical shift information to help populate the Biological Magnetic Resonance Data Bank [71].…”

NMR chemical shift assignments of RNA oligonucleotides to expand the RNA chemical shift database

“…NMRFx Analyst (Marchant et al, 2019). Similarly, these additional data will aid in machine learning approaches that predict RNA secondary structure (Zhang and Frank, 2020) and other RNA structural features from the assigned chemical shift data (Zhang et al, 2021). 16…”