CROSSalive: a web server for predicting the in vivo structure of RNA molecules

Abstract

Here we introduce CROSSalive, an algorithm to predict the RNA secondary structure profile (double and single stranded regions) in vivo and without sequence length limitations. Using predictions of protein interactions CROSSalive predicts the effect of N6 adenosine methylation (m6a) on RNA structure. Trained on icSHAPE data in presence (m6a+) and absence (m6a-) of methylation CROSSalive achieves an accuracy of 0.88 on the test set. The algorithm was also applied to the murine long non-coding RNA Xist (17900 nt, not present in the training) and shows a Pearsons correlation of 0.45 with SHAPE-map data. CROSSalive webserver is freely accessible at the following page: http://service.tartaglialab.com/new_submission/crossalive