Protein Binding FEP+ Benchmark

Data and analysis scripts for Sampson, J. M. et al. "Robust prediction of relative binding energies for protein-protein complex mutations using free energy perturbation calculations." J. Mol. Biol. 436, 168640 (2024). https://doi.org/10.1016/j.jmb.2024.168640

The following analysis was performed on the systems in the paper.

1. Prepare and process the experimental data, downloading the SKEMPI2 data file if necessary.

   ( cd experimental_data && ./prepare_data.sh )

2. Place original 100 ns out.fmp files into results/fmp. Due to size limitations, these files are not included in this repository.

3. Edit the results/systems.csv file to include the relevant metadata for each system.

4. Compile the FEP results dataset, including experimental data and derived data columns.

   ( cd results && ./compile_dataset.sh -v )

5. Flag probable outliers from the .fmp files.

   export SCHRODINGER=/path/to/Schrodinger_Suite_2024-2>
   (
       cd results
   
       # Calculate the .fmp features (this may take a while!)
       time ./calculate_features.sh -v
   
       # Merge the features and results into a single dataset for stats
       Rscript merge_features_and_results.R
   
       # Do the same for testing the case studies dataset
       Rscript merge_features_and_results.R --case-studies
   
       # Flag probable outliers in the benchmark dataset
       # (writes `charged_flag_correction_term.txt`)
       Rscript flag_outliers.R \
           benchmark_merged_features_results.csv \
           -o benchmark_flagged_cases.csv
   
       # Flag probable outliers in the case study dataset using
       # `charged_flag_correction_term.txt` from the benchmark dataset
       Rscript flag_outliers.R \
           case_studies_merged_features_results.csv \
           -o case_studies_flagged_cases.csv
   )