ChemBounce

Dependencies

Install conda environment as follows:

conda env create -n ENVIRONMENT_NAME -f environment.yml

For download reference files,

bash install.sh

How to run

Before running, make sure conda environment is activate conda activate ENVIRONMENT_NAME

python chembounce.py -o OUTPUT_DIRECTORY -i INPUT_SMILES -n FRAGMENT_MAX_N -t TANIMOTO_THRESHOLD

See the example of run.sh

Parameters

-i INPUT_SMILES : Input SMILES, the target molecular structure
-o OUTPUT_DIRECTORY : Output location
-n FRAGMENT_MAX_N : (int) "Maximal number of scaffold-hopped candidates for a fragment
-t TANIMOTO_THRESHOLD : (float; 0 to 1) Threshold for Tanimoto simiarity between INPUT_SMILES and generated SMILES. (default is 0.5)

Options for iteration numbers

Limit numbers of iterations

--overall_max_n : Maximal number of scaffold-hopped candidates for overall fragments
--scaffold_top_n : Number of scaffolds to test for a fragment
--cand_max_n__rplc : Maximal number of candidates for a replaced scaffold

Options for Thresholds

Thresholds for candidates

-t TANIMOTO_THRESHOLD : (float; 0 to 1) Threshold for Tanimoto simiarity between INPUT_SMILES and generated SMILES. (default is 0.5)
--qed_min QED_MIN : (float) Minimal QED score
--qed_max QED_MAX : (float) Maximal QED score
--sa_min SA_MIN : (float; 0.0 to 10.0) Minimal SAscore
--sa_max SA_MAX : (float) Maximal SAscore
--logp_min LOGP_MIN : (float) Minimal logP limit
--logp_max LOGP_MAX : (float) Maximal logP limit
--mw_min MW_MIN : (float; above 0.0) Minimal molecular weight
--mw_max MW_MAX : (float; above 0.0) Maximal molecular weight
--h_donor_min H_DONOR_MIN : (above 0) Minimal H donor limit
--h_donor_max H_DONOR_MAX : (above 0) Maximal H donor limit
--h_acceptor_min H_ACCEPTOR_MIN : (above 0) Minimal H acceptor limit
--h_acceptor_max H_ACCEPTOR_MAX : (above 0) Maximal H acceptor limit
--wo_lipinski : Turn off Lipinski's rule of five, which is logp_max=5, qed_max=500, h_donor_max=5, h_acceptor_max=10 This option will turn off the limitation of candidates by lipinski's rule of five If one of the threhold category for this rule is additionally defined, the threshold of Lipinski's rule is ignored and replaced by the user-defined threshold.

Optional parameters

-l : (Optional) Low memory mode
--fragments : Specific fragment SMILES of the input structure to replace. For multiple fragment, repeatedly impose this option: e.g. --fragments SMILES_A --fragments SMILES_B
--replace_scaffold_files : Replace scaffold file, for the fragments option with tsv format, with or without priority score (the higher, with more priority).
- NOTE: One of the result files, replace_scaffold_list.fragment_XXXX.tsv, can be used for replace_scaffold_files.
- NOTE: for multiple fragments and its corresponding replace_scaffold_files, the order of fragments and replace_scaffold_files should be matched: e.g.: --fragments SMILES_A --replace_scaffold_files FILE_A --fragments SMILES_B --replace_scaffold_files FILE_B --fragments SMILES_C --replace_scaffold_files FILE_C
- NOTE: For multiple

Results

Result file will is saved as OUTPUT_DIRECTORY/overall_result.txt with tap-delimitted format. Final structure column is the final result SMILES of scaffold hopping and Standardized final structure is standardized SMILES format for the corresponding result. For sub-data,

Fragments of the input molecule is saved as ~/fragment_info.tsv
Considered replace scaffold list for each fragment is saved as ~/replace_scaffold_list.fragment_XXXX.tsv.
Results for each fragment is saved ~/fragment_result_XXXX.txt.

Citation

Contact

For more information : check CSB_Lab

jyryu3161/chembounce