BioCatalyzer is a python tool that predicts enzymatic metabolism products using a rule-based approach.

BioCatalyzer is implemented as a Command Line Interface that takes as input a set of compounds represented as SMILES strings and outputs a set of predicted metabolic products and associated enzymes.

This metabolic products can then be matched with experimental MS data using this same tool.

Installing from Pypi package repository:

pip install biocatalyzer

Installing from GitHub:

1. clone the repository: git clone https://github.com/jcorreia11/BioCatalyzer.git

2. run: python setup.py install

biocatalyzer_cli <PATH_TO_COMPOUNDS> <OUTPUT_DIRECTORY> [--neutralize=<BOOL>] [--reaction_rules=<FILE_PATH>] [--organisms=<FILE_PATH>] [--patterns_to_remove=<FILE_PATH>] [--molecules_to_remove=<FILE_PATH>] [--min_atom_count=<INT>] [--match_ms_data=<BOOL>] [--ms_data_path=<FILE_PATH>] [--tolerance=<FLOAT>] [--n_jobs=<INT>]

See drugs.csv¹ for an example.

The file must be tab-separated and contain the following columns:

• smiles - the SMILES representation of the compounds;
• compound_id - the compounds identifiers.

Alternatively, the compounds can be passed as ;-separated string with the SMILES representations.

The output path must be a directory. The results will be saved in the following files:

• new_compouds.tsv - the predicted products;
• matches.tsv (if match_ms_data is set to True) - the matches between the predicted products and the MS data;

If set to True, the compounds will be neutralized before predicting the products. In this case the new products will also be neutralized.

See all_reaction_rules_forward_no_smarts_duplicates_sample.tsv² for an example.

The file must be tab-separated and contain the following columns:

• InternalID - The ID of the Reaction Rule. # TODO: change the name of this column
• Reactants - The Reactants of the ReactionRule. Coreactants must be defined by their ID as in the Coreactants file. The compound to match must be identified by the string 'Any'. The format must be: coreactant1_id;Any;coreactant_id. The order in which the reactants and the compound to match are defined is relevant and specific to the Reaction Rule. If the Reaction Rules are mono-component (i.e. they do not contain any additional coreactant) the format must be: Any.
• SMARTS - The SMARTS representation of the Reaction Rule.
• EC_Numbers - The EC Numbers associated with the Reaction Rule.
• Organisms - The Organisms associated with the Reaction Rule.

By default our set of reaction rules is used.

All organisms' identifiers are defined in: https://www.genome.jp/kegg/catalog/org_list.html are allowed.

Example:

hsa is for Homo sapiens (human).

eco is for Escherichia coli K-12 MG1655.

sce is for Saccharomyces cerevisiae (budding yeast).

If you want to use your own organisms see organisms.csv³ for an example.

The file must be tab-separated and contain a column named org_id with the organisms' identifiers (KEGG identifiers).

Alternatively, the organisms can be passed as ;-separated string with the organisms identifiers.

If you want to use your own patterns to remove see patterns.tsv⁴ for an example.

The file must be tab-separated and contain a column named smarts with the SMARTS representation of the patterns to remove.

If you want to use your own molecules to remove see byproducts.tsv⁵ for an example.

The file must be tab-separated and contain a column named smiles with the SMILES representation of the molecules to remove.

If set to True, the predicted products will be matched to the MS data.

In this case the ms_data_path must be set.

See ms_data.tsv⁶ for an example.

The file must be tab-separated and contain the following columns:

• ParentCompound - the parent/original compound identifiers.
• ParentCompoundSmiles - the SMILES representation of the compounds (optional).
• Mass - the mass of the molecule.

The mass tolerance (float) to use when matching masses. Masses between mass - mass_tolerance and mass + mass_tolerance will be considered as a match.

The number of jobs to run in parallel. If -1 is passed, all available cores will be used.

biocatalyzer_cli file.tsv output_dir/ --neutralize=True --reaction_rules=reaction_rules.tsv --organisms="hsa;eco;sce" --patterns_to_remove=patterns.tsv --molecules_to_remove=byproducts.tsv --match_ms_data=True --ms_data_path=ms_data.tsv --mass_tolerance=0.1 --n_jobs=-1

For predicting compound metabolism only:

biocatalyzer_cli file.tsv output_dir/ --neutralize=True --reaction_rules=reaction_rules.tsv --organisms="hsa;eco;sce" --patterns_to_remove=patterns.tsv --molecules_to_remove=byproducts.tsv --n_jobs=-1

Both parts of this CLI (the generation of new compounds (bioreactor_cli) and the matching with the MS data (matcher_cli)) can be run individually.

For the bioreactor_cli see readme_bioreactor_cli.md.

For the matcher_cli see readme_matcher_cli.md.

Manuscript under preparation!

Developed at Centre of Biological Engineering, University of Minho and EMBL Heidelberg (Zimmermann-Kogadeeva Group).

This project has received funding from the Portuguese FCT and EMBL CPP Scientific Visitors Fellowships.

Released under an MIT License.