A high-throughput approach to profile RNA structure

Delli Ponti, Riccardo; Marti, Stefanie; Armaos, Alexandros; Tartaglia, Gian Gaetano

The CROSS algorithm CROSS predicts the structural profile of an RNA sequence at single-nucleotide resolution and without sequence length restrictions. The algorithm is an artificial neural network with one hidden layer and two adaptive weight matrices to predict the structural state of a nucleotide considering its flanking residues (Materials and Methods: CROSS architecture; Supplementary Material: Selection of the optimal window). We built five independent models using data from SHAPE ( either single-stranded (negative cases) or double-stranded (positive cases) configuration. Each model was then tested on all the other data sets. Negligible overlap exists between training and testing sets (Jaccard index < 0.002 between each couple of sets analyzed; Supplementary Table S5) .