PD-1 Targeted Antibody Discovery Using AI Protein Diffusion

Diffusion of Novel, PD-1-targeting Fab Structures with in silico Evaluation

Tuple, LLC

Antibody Generation and Testing

Steps:

0. Generate Sequences using EvoDiff
   • Follow EvoDiff logic in scripts/0_sequence_generation/EvoDiff_sequence_generation.md
1. Generate Structures using AlphaFold2 (TUPPD1-001 to TUPPD1-009)
   • Run AlphaFold2 Batch using scripts/1_structure_generation/AlphaFold2_batch.ipynb
2. Prepare Structures
   • Renumber L chains in PDB files with PyMol: scripts/2_structure_prep/PDB_renumberer.ipynb
   • Detect CDR loop residues: scripts/2_structure_prep/fab_cdr_detection.ipynb
3. Generate HADDOCK experiment files
   • Run experiment generation using scripts/3_experiment_generation/make_HADDOCK_experiments.ipynb
4. Submit HADDOCK experiments to HPC
   • Create Singularity container from Docker image: singularity build haddock.sif docker://cford38/haddock:2.4_36cores
   • Run experiment generation: bash ./scripts/4_experiment_submission/submit_experiments.sh
     • (Remember to chmod -R 755 the experiments/ folder so that Singularity can execute the scripts.)
5. Collect best docked PDB structure and metrics for each experiment
   • Run experiment metrics collection: scripts/5_metrics_collection/postprocessing.ipynb
     • This notebook also runs PRODIGY to calculate binding affinities for each structure.
6. Clean up HADDOCK run files
   • Run cleanup to move each experiments' run1/ directory into outputs: scripts/6_cleanup/cleanup.ipynb X. Run various post-analyses:
   • abYsis for analyzing the residues in the Fabs: analysis/abysis
   • OpenMM for molecular dynamics of the docked structures: analysis/molecular_dynamics/run_openmm.ipynb