Output and Extraction

Output files of any of the prediction methods (sampling, clote-ponty, pmcomp) contain just a Python dictionary of sequence identifiers as keys and numpy arrays as values. It is possible to load it via pythons pickle module. You can access the expected distance between nucleotide i and j via their corresponding indices in this array. For example the distance between the first nucleotide and the 20th can be accessed like this:

import pickle
with open(output_file, "rb") as handle:
        rnadist_sample_output: Dict[str, np.ndarray] = pickle.load(handle)
ed1_20 = rnadist_sample_output[0][19]

Warning

In contrast to ViennRNA this Tool uses 0 based indexing

However, if you want to use another programming Language for further work, the pickled expected distance files can be extracted to TSV files via the command line. This is done via the following call:

RNAdist extract --data_file path/to/predict/output.pckl --outdir example_dir

This will produce a single tsv file for each sequence that was used in the prediction step. All tsv files will be saved in the outdir and their names will be equal to the FASTA sequence identifiers. For further understanding have a look at the RNAdist Command Line Interface.