enzyme_screen

Scripts and functions for extracting and analysing biochemical reactions.

Author: Andrew Tarzia Email: andrew.tarzia@gmail.com or atarzia@ic.ac.uk

This work was produced in the final year of my PhD at the University of Adelaide under the supervision of A/Prof David Huang and Prof Christian Doonan.

Previously at: https://bitbucket.org/andrewtarzia/psp_source/src/master/

A Jupyter notebook that runs through the molecular size calculation from a SMILES string is available:

The molecular size calculation code is also available in a refactored form at my GitHub and through PyPi: https://github.com/andrewtarzia/mol-ellipsize

Installation

• Tested on Ubuntu 18.04 using conda and pip
• Install Anaconda in standard way (Python 3.7.3)
• Packages required outside of what comes with conda
• RDKit:
  • conda install -c conda-forge rdkit
  • Version: 2019.09.2.0
• chemcost:
  • Python code written by Steven Bennett for the extraction of purchasability from the ZINC15 database.
  • Follow instructions found here: https://github.com/stevenbennett96/chemcost
  • Only required for molecule_population.py

Workflow

Collecting database from KEGG

• Download br08201 JSON file from the KEGG library
• Used version as of May12_2020 of br08201: Enyzmatic reactions
• Run util/split_KEGG.py in working directory to produce:
  • _ECtop.json: A dictionary of all reactions for all ECs
  • _EClist.txt: A list of all ECs to iterating through
• Update data/param_file.txt with location of these files.

Parameter testing

• All parameter screens in the supporting information of DOI: awaiting are run in param_screening.py
  • data/test_molecules.txt contains the required molecular information
  • Within param_screening.py are the range of parameters to test, the originals being set in data/param_file.txt

Reaction collection and analysis

• RS_collection.py

  • Iterates through provided EC and reaction files to collect reaction systems
  • Also collects unique molecules to molecule database
  • Currently only implements API for KEGG
  • To be run in directory with reactions

• molecule_population.py

  • Trivial parallelisation done using utils/molecule_splitter.py
  • Takes _unopt.mol file of all collected molecules:
    • Optimises them using ETKDG -> _opt.mol
    • Calculates their properties -> _prop.json
    • Calculates the molecule size of N conformers -> _size.csv
  • To be run in directory with molecules
  • Produces some plots of chemical space

• chemical_space_plot.py

  • Iterates through all collected molecules and plots various chemical space plots
  • To be run in directory with molecules

• RS_analysis.py

  • molecule_population.py must be run before this point!
    • Unanalysed molecules result in skipped reactions
  • Populates the properties of each reaction system based on the properties of constituent molecules (in molecule database)
  • To be run in directory with reactions
  • Outputs all properties to rs_properties.csv

• screening.py

  • Produces the plots and screening of all reaction systems seen in DOI: awaiting
  • Multiple cases are defined within the script to look at specific EC numbers or system types
    • case = production for plots in DOI:
  • To be run in directory with reactions

Examples

• biomin_screening.py

  • A script used to produce Figure XX in DOI:
  • Analyses a list of molecules that have been tested for enzyme@ZIF-8 reactions

• examples/calculate_molecular_size.ipynb

  • Jupyter notebook that runs a user through calculating the size of any molecule

• examples/screen_new_reactions.ipynb

  • Jupyter notebook that runs through the screening process exemplified in the paper search for new reactions

• visualise_ellipsoid_steps.py

  • Allows the user to visualise the step-wise calculation of the min. vol. enclosing ellipsoid

• visualise_reaction_system.py

  • Allows the user to print properties of a reaction system