Closed alyssadai closed 2 years ago
Thanks for the changes. In https://github.com/CoBrALab/cobra-nmf/pull/15/commits/7da36ae70342e9b57ab5bf9e9182b727f294af2f, can you walk me through whats happening in line 147?
resid_vars_a = resid_vars[Asplits_indices[str(split)],:]; resid_vars_b = resid_vars[Bsplits_indices[str(split)],:]
following the code i thought resid_vars
is, at this point, a list containing the variables to residualize for, but its being indexed above more like it contains data for each subject? Assuming i'm missing something, lemme know.
Nope, you're right, that's an error on my part. I had the names of residualized variables hard coded in my original copy of the script, and accidentally got rid of the variable storing the corresponding columns from the demographics file when switching to the command-line argument. Should be fixed in https://github.com/CoBrALab/cobra-nmf/pull/15/commits/99c9110ae8baa168cba87dea8655a6dfb1629e28. Thanks for catching that!
Added optional
--output_suffix
argument in extract_metrics.py (accepts a string) and--residfor
argument in define_splits.py (accepts a list).Users now have more flexibility in naming the single-metric input files (addresses #14), and can run stability analyses where the vertex input in splits are residualized for specified variables from the demographics sheet.