Predicting RNA Secondary Structure Using In Vitro and In Vivo Data.

Abstract

The new flow of high-throughput RNA secondary structure data coming from different techniques allowed the further development of machine learning approaches. We developed CROSS and CROSSalive, two algorithms trained on experimental data able to predict the RNA secondary structure propensity both in vitro and in vivo. Since the in vivo folding of RNA molecules depends on multiple factors due to the cellular crowded environment, prediction is a complex problem that needs additional calculations for the interaction with proteins and other molecules. In the following chapter, we will describe the differences in predicting RNA secondary structure propensity using experimental data as input for an Artificial Neural Network (ANN) in vitro and in vivo.