Update input/output file system structure

[dominiquesydow]

Description

IO file system

Unify input/output file system structure between the KinFragLib and KinaseFocusedFragmentLibrary, so that SDF files are grouped in folders by library type instead of by subpocket.

File system structure in KinaseFocusedFragmentLibrary

FragmentLibrary/AP/AP.sdf
FragmentLibrary/AP/AP_reduced_0.6.sdf

FragmentLibrary/SE/SE.sdf
FragmentLibrary/SE/SE_reduced_0.6.sdf

# and so on

File system structure in KinFragLib

data/fragment_library/AP.sdf
data/fragment_library/SE.sdf
# and so on

data/fragment_library_reduced/AP.sdf
data/fragment_library_reduced/SE.sdf
# and so on

Combinatorial library output type

Currently, the final combinatorial library (molecule meta data + analysis) is stored as pickle file which is not OS and library version independent. Save as json instead.

Todos

• [x] Remove RDKit version restriction from environment.yml file
• [x] Update IO folder structure in
  • [x] fragmentation/cli.py
  • [x] recombination/cli.py
  • [x] analysis/ligand_analysis.py
• [x] Update README to account for new IO; add remark for notebooks (not all will run out-of-the-box with new IO - but also before most of them needed path updates)
• [x] Save final combinatorial library (after ligand analysis) as json file (instead of pickle)
• [x] Fixes in ligand analysis scripts
  • [x] Fix error for substructure match in original ligands: Use HasStubstructMatch instead of >= comparison.
  • [x] Where does n_subpockets come from? Not needed, since this information lives in meta.frag_ids, right?

Questions

None.

Status

• [x] Ready to go

[dominiquesydow]

This is the code version used to generate the KinFragLib paper results (https://chemrxiv.org/s/2c0fb3a743dec8994fc0).

Merge back to master and tag this code version as 1.0.0.