Web demos sincFold

Resources

Source code and data: (1) | Related publications: [1]

sincFold is an end-to-end deep learning approach that predicts the nucleotides contact matrix using only the RNA sequence as input. This demo provides a sincFold model pre-trained with known RNAs secondary structures.

Input RNA sequences of up to 512 nucleotides can be processed with this demo. To process sequences longer than 512 nucleotides you can either install the package (via pip) or clone the entire repository.

Sample inputs:

  • AACCGGGUCAGGUCCGGAAGGAAGCAGCCCUAA
  • CCACGGCGACUAUAUCCCUGGUGUUCACCUCUUCCCAUUCCGAACAGAGUCGUUAAGCCCAGGAGAGCCGAUG
    GUACUGCUUUAUUGCGGGAGAGUAGGUCGUCGCCGAGU
  • GAUAAACCUUUAGCAAUAAACGAAAGUUUAACUAAGCCAUACUAACCCCAGGGUUGGUCAAUUUCGUGCCAGC
    CACCGCGGUCACACGAUUAACCCAAGCCAAUAGAAAUCGGCGUAAAGAGUGUUUUAGAUCAAUCCCCCAAUAA
    AGCUAAAAUUCACCUG

To process multiple sequences at once or retrain the model, please follow the instructions provided with the source code.

 

Contact: Leandro Bugnon

 

* Conversion from the CT output of sincFold to dot-bracket format and svg are done with RNAstructure (it can fail with pseudoknots).
** The interactive structure was generated from the CT output of sincFold with forna.