Current life on earth is believed to have descended from self-replicating RNA molecules, and evidence of its deep history is apparent in its diversity of function. Discerning the structure of RNAs will help inform our understanding of RNA function as well as develop future tools in medicine. The most basic structural information lies in its secondary structure, the first level of structural organization within an RNA molecule. Techniques such as crystallography, comparative analysis, and computational algorithms have been developed to predict secondary structure of RNA, though performance becomes hindered when analyzing longer RNAs. In order to find a more efficient method of predicting the structure of long RNAs, we combine data gathered from DMS probing experiments (Structure-Seq) and input subsections of the RNA into the RNAStructure prediction algorithm. By dividing the structure into smaller sections, we find that predictive capabilities can be vastly improved, though inclusion of DMS probing data has varying effects in improving prediction accuracy. We test this subdividing of RNA of prediction in both a user-directed and naive manner in the 18S RNA in Arabidopsis thaliana. Overall, these improvements in computation and experimentation suggest a more efficient and accurate strategy to predict RNA secondary structure in long RNAs.
Structure-Seq Influenced RNA Secondary Structure Prediction