The 2dRNA web server provides services for predicting RNA secondary structures including a wide class of pseudoknots. 2dRNA is an end-to-end deep learning-based method, which combines two famous neural network architecture bidirectional-LSTM and U-net. 2dRNA could use both the structural knowledge and evolutionary information learning from similar RNA sequences to predict the secondary structure of a RNA. 2dRNA is trained on a large dataset, so after training it can runs fast for prediction, and the only needed input is just a sequence.
Data sets downloads
The data used in our paper can be downloaded below, you can download them and validate them in our web server. Please read the README file to understand how the data is structured.