This repository contains the code and data used for the article 'Improving RNA branching predictions: advances and limitations' by Svetlana Poznanovic ́, Carson Wood, Michael Cloer, and Christine E. Heitsch (link will be made available).
These scripts are written for SageMath in accompaniement with Python 3. They are known to work with SageMath 9.2 and Python 3.7, although they likely are functional with earlier versions too.
To run, use the bb_main.sage script. For example:
sage ~/bb_main.sage
This will find the optimal parameters for the sequences listed in the file specified as seq_list in the main code.
The optimal parameters are the parameters achieving the best average accuracy value across the specified data set.
The region containing those parameters will be specified in ~/MergeData/FinalOutput.txt.
Additionally, data about algorithm computations and runtimes will be listed in ~/MergeData/SummativeData.txt.
We are providing input data for 50 tRNA and 50 5S rRNA sequences that were used for training. The data files include the primary sequences, the pseudoknot-free secondary structures, the branching polytopes in the .rnapoly files computed using the pmfe, and the optimal structures computed for all vertices of the branching polytopes.