How to run inference with a method ensemble¶
python3 multicom_ligand/models/ensemble_generation.py ensemble_methods=[tulip] input_csv_filepath=data/test_cases/casp15/ensemble_inputs.csv output_dir=data/test_cases/casp15/top_tulip_ensemble_predictions_1 skip_existing=true relax_method_ligands_post_ranking=false export_file_format=casp15 export_top_n=5 combine_casp_output_files=true max_method_predictions=40 method_top_n_to_select=40 resume=true ensemble_benchmarking=true ensemble_benchmarking_dataset=casp15 cuda_device_index=0 ensemble_benchmarking_repeat_index=1 python3 multicom_ligand/models/ensemble_generation.py ensemble_methods=[tulip] input_csv_filepath=data/test_cases/casp15/ensemble_inputs.csv output_dir=data/test_cases/casp15/top_tulip_ensemble_predictions_1 skip_existing=true relax_method_ligands_post_ranking=true export_file_format=casp15 export_top_n=5 combine_casp_output_files=true max_method_predictions=40 method_top_n_to_select=40 resume=true ensemble_benchmarking=true ensemble_benchmarking_dataset=casp15 cuda_device_index=0 ensemble_benchmarking_repeat_index=1
NOTE: the suffixes for both output_dir and ensemble_benchmarking_repeat_index should be modified to e.g., 2, 3, …¶
… now score the CASP15-compliant submissions using the official CASP scoring pipeline ===================================================================================
python3 multicom_ligand/analysis/inference_analysis_casp.py method=tulip dataset=casp15 targets=’[H1135, H1171v1, H1171v2, H1172v1, H1172v2, H1172v3, H1172v4, T1124, T1127v2, T1152, T1158v1, T1158v2, T1158v3, T1158v4, T1186, T1187]’ …
</details>
## How to run inference with a method ensemble
<details>
Using an `ensemble` of methods, generate predictions for a new protein target using each method and (e.g., consensus-)rank the pool of predictions (n.b., see the function `execute_steps` within `scripts/execute_casp16_ensemble_generation_strategy.py` for more details regarding `MULTICOM_ligand`'s usage during CASP16)
```bash
# generate each method's prediction script for a target
# NOTE: to predict input ESMFold protein structures when they are not already locally available in `data/ensemble_proteins/`, e.g., on a SLURM cluster first run e.g., `srun --partition=gpu --gres=gpu:A100:1 --mem=59G --time=01:00:00 --pty bash` to ensure a GPU is available for inference
python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/5S8I_2LY/ensemble_inputs.csv output_dir=data/test_cases/5S8I_2LY/top_consensus_ensemble_predictions_1 max_method_predictions=40 ensemble_ranking_method=consensus resume=false ensemble_methods='[diffdock, dynamicbind, neuralplexer, rfaa]'
# ...
# now, manually run each desired method's generated prediction script, with the exception of AutoDock Vina which uses other methods' predictions
# ...
python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/5S8I_2LY/ensemble_inputs.csv output_dir=data/test_cases/5S8I_2LY/top_consensus_ensemble_predictions_1 max_method_predictions=40 ensemble_ranking_method=consensus resume=true generate_vina_scripts=true vina_binding_site_methods=[diffdock]
# now, manually run AutoDock Vina's generated prediction script for each binding site prediction method
#...
# lastly, organize each method's predictions together
python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/5S8I_2LY/ensemble_inputs.csv output_dir=data/test_cases/5S8I_2LY/top_consensus_ensemble_predictions_1 max_method_predictions=40 ensemble_ranking_method=consensus resume=true generate_vina_scripts=false vina_binding_site_methods=[diffdock]
Benchmark (ensemble-)ranked predictions across each test dataset
benchmark using the PoseBusters Benchmark dataset e.g., after generating 40 complexes per target with each method¶
python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/posebusters_benchmark/ensemble_inputs.csv output_dir=data/test_cases/posebusters_benchmark/top_consensus_ensemble_predictions_1 max_method_predictions=40 export_top_n=1 export_file_format=null skip_existing=true relax_method_ligands_post_ranking=false resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, tulip, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=posebusters_benchmark ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/posebusters_benchmark/ensemble_inputs.csv output_dir=data/test_cases/posebusters_benchmark/top_consensus_ensemble_predictions_1 max_method_predictions=40 export_top_n=1 export_file_format=null skip_existing=true relax_method_ligands_post_ranking=true resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, tulip, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=posebusters_benchmark ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 …
benchmark using the Astex Diverse dataset e.g., after generating 40 complexes per target with each method¶
python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/astex_diverse/ensemble_inputs.csv output_dir=data/test_cases/astex_diverse/top_consensus_ensemble_predictions_1 max_method_predictions=40 export_top_n=1 export_file_format=null skip_existing=true relax_method_ligands_post_ranking=false resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, tulip, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=astex_diverse ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/astex_diverse/ensemble_inputs.csv output_dir=data/test_cases/astex_diverse/top_consensus_ensemble_predictions_1 max_method_predictions=40 export_top_n=1 export_file_format=null skip_existing=true relax_method_ligands_post_ranking=true resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, tulip, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=astex_diverse ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 …
benchmark using the DockGen dataset e.g., after generating 40 complexes per target with each method¶
python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/dockgen/ensemble_inputs.csv output_dir=data/test_cases/dockgen/top_consensus_ensemble_predictions_1 max_method_predictions=40 export_top_n=1 export_file_format=null skip_existing=true relax_method_ligands_post_ranking=false resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=dockgen ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/dockgen/ensemble_inputs.csv output_dir=data/test_cases/dockgen/top_consensus_ensemble_predictions_1 max_method_predictions=40 export_top_n=1 export_file_format=null skip_existing=true relax_method_ligands_post_ranking=true resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=dockgen ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 …
benchmark using the CASP15 dataset e.g., after generating 40 complexes per target with each method¶
python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/casp15/ensemble_inputs.csv output_dir=data/test_cases/casp15/top_consensus_ensemble_predictions_1 combine_casp_output_files=true max_method_predictions=40 export_top_n=5 export_file_format=casp15 skip_existing=true relax_method_ligands_post_ranking=false resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, tulip, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=casp15 ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 python3 multicom_ligand/models/ensemble_generation.py input_csv_filepath=data/test_cases/casp15/ensemble_inputs.csv output_dir=data/test_cases/casp15/top_consensus_ensemble_predictions_1 combine_casp_output_files=true max_method_predictions=40 export_top_n=5 export_file_format=casp15 skip_existing=true relax_method_ligands_post_ranking=true resume=true cuda_device_index=0 ensemble_methods=’[diffdock, dynamicbind, neuralplexer, rfaa, tulip, vina]’ ensemble_benchmarking=true ensemble_benchmarking_dataset=casp15 ensemble_ranking_method=consensus ensemble_benchmarking_repeat_index=1 …
analyze benchmarking results for the PoseBusters Benchmark dataset¶
Note
In addition to having consensus as an available value for ensemble_ranking_method, one can also set ensemble_ranking_method=ff to have the method ensemble’s top-ranked predictions selected using the criterion of “minimum (molecular dynamics) force field energy” (albeit while incurring a very large runtime complexity).