Skip to content

1_validate_input.md

Latest commit

 

History

History
39 lines (28 loc) · 1.43 KB

1_validate_input.md

File metadata and controls

39 lines (28 loc) · 1.43 KB
Raw
name Validate input protein [cofactor]
description input validation using OpenFE CLI.

Step 1. Validate input pdb file by running a short MD simulation

Assume input files are these by default: a. ligands.sdf b. protein.pdb c. cofactors.sdf (this file is optional)

This step test if the input protein pdb file has any known problems, for example, invalid names for teriminal capping groups. It does so by running a short MD simulation using the protein file, and the cofactors file if provided.

The command is:

SKILL_DIR="$HOME/.claude/skills/openfe_skill"
export SKILL_DIR
source "$SKILL_DIR/scripts/hpc_config.sh"
sbatch "${SBATCH_INPUT_VALIDATION[@]}" \
    $SKILL_DIR/scripts/run_input_validation_cv.sh protein.pdb cofactors.sdf

All SLURM directives (account, partition, GPU, walltime, memory, output files) come from the SBATCH_INPUT_VALIDATION array in scripts/hpc_config.sh -- edit that file to adapt to a different cluster.

If validation is completed successfully, the job's input_validation_$JOBID.out file should look like this at the end:

Protein file: protein.pdb

SIMULATION COMPLETE

If validation failed, show user the error message(s) from the .err file.

Agent Instructions for this Step:

  1. Do not copy the folder "scripts" to the working folder.
  2. Save or update information about this step, the command line executed and the SLURM job ID in a file named Job.log. Create the file if absent.