Skip to content
Community

Dock, Validate, and Inspect a Protein–Ligand Pose Locally

The workflow was executed on Linux x86_64 CPU using the public 1EVE/E20 example. GNINA 1.3.3 generated three poses, PoseBusters 0.6.5 checked all three, and PLIP 2.4.0 analyzed all three with matching chemical identity. This is computational workflow validation; no experimental or biological validation has been performed.

Level: Advanced Cost: Free Privacy: Local ~60 min

Editorially updated:

Start Setup
Open in WorkBuddy
Third-party service

About WorkBuddy

This feature is provided by WorkBuddy, a third-party service, and is not an official ChemAI Atlas runtime.

Opens an external website. Data processing and service use are governed by WorkBuddy’s own terms.

This experimental app supports setup assistance for some Recipes; direct execution is not guaranteed.

You'll be able to

  • Audit pose files and computational plausibility
  • Preserve interaction-report provenance

What you'll build

The workflow was executed on Linux x86_64 CPU using the public 1EVE/E20 example. GNINA 1.3.3 generated three poses, PoseBusters 0.6.5 checked all three, and PLIP 2.4.0 analyzed all three with matching chemical identity. This is computational workflow validation; no experimental or biological validation has been performed.

Tested host: Ubuntu 24.04.2, Linux x86_64, two vCPUs, about 8 GiB RAM, Python 3.12.15. The project has its own environment and versioned GNINA binary; the SHA-256 is 3340c1f49cd3c7c84d8699182a1c6af13c7fa2a22448d1204640446106f72172. Although execution uses --no_gpu, this binary needs CUDA/cuDNN loader libraries. They are installed only in the project environment and available through the isolated environment, together with the Open Babel Xrender dependency. No system CUDA driver or global Python modification was needed. The measured GNINA command took about 45 seconds on this host; this is one observation, not a runtime guarantee. The recorded installation and shared caches occupied about 8.3 GiB; this is an observed footprint, not a minimum storage requirement. Downloads require network access; no privacy audit is claimed.

Docking/CNN scores are not measured affinity, IC50 or efficacy. PoseBusters checks configured plausibility and reference criteria; PLIP reports rule-based contacts in supplied coordinates. This one known-pocket fixture does not establish general docking accuracy, cross-platform compatibility or biological activity. No independent prediction RMSD calculation was performed: the report retains PoseBusters' configured RMSD-threshold booleans. Neutral ligand choice, stripped water/glycans and unmodeled termini constrain interpretation.

Setup time is an estimate, not a measured total installation time.

gnina

GNINA

posebusters

PoseBusters

plip

PLIP

Stack Components

GNINA

gnina · 1.3.3; upstream binary SHA-256 pinned

Executed on Linux x86_64 CPU.

View Resource

PoseBusters

posebusters · 0.6.5

Executed on Linux x86_64 CPU.

View Resource

Compatibility

ClientOSArchitectureVersion requirements
Python SupportedLinuxx86_64>= 3.12

Setup & Test

1. Check the controlled project environment

Linux

The commands below retain the parameters of the recorded calculation, with example paths. Replace /path/to/molecular-docking-validation with your own directory. Prerequisites are installed tools with their runtime libraries available on PATH and the prepared receptor.pdb, independently generated ligand.sdf, and coordinate-matched reference.sdf in data/prepared/. PoseBusters additionally needs its installed redock.yml copied to runs/example-run/config/redock.yml; PLIP needs each receptor–pose PDB complex already assembled and checked for atom serials, TER/CONECT records, bond orders, coordinates and chemical identity. This article does not distribute a complete project, prepared data or assembly scripts, so these are conditional command examples rather than a self-contained copy-and-run package. Use fresh output directories and preserve the input files, parameters and logs. Execution was validated only on Linux x86_64 CPU with two vCPUs and about 8 GiB RAM; GPU and native macOS/Windows compatibility were not validated.

Do not overwrite the recorded run or environment; use fresh versioned paths.

PROJECT_DIR=/path/to/molecular-docking-validation
RUN_DIR="$PROJECT_DIR/runs/example-run"
INPUT_DIR="$PROJECT_DIR/data/prepared"
cd "$PROJECT_DIR"
mkdir -p "$RUN_DIR"/{dock,config,validation,complex,interactions}
python --version
gnina --version
bust --version
plip --help
Official source

Expected result

Versions captured

2. Provision the isolated tools and loader libraries

Linux

Install GNINA 1.3.3, PoseBusters 0.6.5, PLIP 2.4.0, PDBFixer 1.12.0 and OpenMM 8.6.1 in an isolated environment using their official installation instructions. Python 3.12.15 was used in the recorded calculation. Check the GNINA binary digest against the value in the description and make the tools available on PATH. This binary required CUDA/cuDNN loader libraries even with --no_gpu; PLIP/Open Babel also required Xrender. Missing libraries caused the initial dependency checks to fail and were resolved within the project environment without a system CUDA driver or global Python changes. A nonzero dependency check means the environment still needs repair. No provisioning scripts or complete environment bundle are distributed here.

Do not overwrite the recorded run or environment; use fresh versioned paths.

Official source

Expected result

Environment dependency checks pass after repair

3. Prepare the fixed public fixture

Linux

Use the public RCSB PDB 1EVE structure and CCD E20 ligand definition. The recorded calculation reused previously retrieved source files whose identity was checked by SHA-256; it did not fetch a new copy during execution. Retrieve and identify your own source files, retain provenance and hashes, and prepare receptor.pdb, ligand.sdf and reference.sdf under data/prepared/. Prepared files and a preparation script are not distributed with this article. The receptor uses the observed A chain of 1EVE. All HETATM, including E20, NAG/glycans and water, were removed. PDBFixer 1.12.0 repaired eight side-chain heavy atoms in ASP A:2 and HIS A:3 plus THR A:535 OXT; 4263 heavy atoms remain and none of the original heavy-atom coordinates moved. Unobserved terminal residues were not invented. OpenMM 8.6.1 added receptor hydrogens with pH 7.4 rules on the Reference platform. E20 remains the neutral CCD identity, without a physiological microstate claim. The reference is E20:A:2001 with 28 heavy atoms and InChIKey ADEBPBSSDYVVLD-HXUWFJFHSA-N. The docking input was independently embedded from the isomeric graph using ETKDGv3 seed 20261019 and MMFF, not copied from the crystal pose.

Do not overwrite the recorded run or environment; use fresh versioned paths.

Official source

Expected result

Three prepared inputs and no remaining missing heavy atoms

4. Generate real poses with GNINA

Linux

Use a new output directory; RUN_DIR is an example path. Preserve stdout/stderr and command parameters. The executed command generated three finite 3D SDF records with matching InChIKey and observed minimizedAffinity, CNNscore, CNNaffinity, CNN_VS and CNNaffinity_variance properties. Do not infer measured affinity from those names.

Do not overwrite the recorded run or environment; use fresh versioned paths.

PROJECT_DIR=/path/to/molecular-docking-validation
RUN_DIR="$PROJECT_DIR/runs/example-run"
INPUT_DIR="$PROJECT_DIR/data/prepared"
cd "$PROJECT_DIR"
mkdir -p "$RUN_DIR"/{dock,config,validation,complex,interactions}
gnina -r "$INPUT_DIR/receptor.pdb" -l "$INPUT_DIR/ligand.sdf" \
  --autobox_ligand "$INPUT_DIR/reference.sdf" --seed 20261019 \
  --num_modes 3 --exhaustiveness 8 --cpu 2 --no_gpu \
  --cnn_scoring rescore -o "$RUN_DIR/dock/poses.sdf"
Official source

Expected result

Three parsed GNINA poses

Actual negative runs covered missing/invalid receptor, empty/invalid SDF, missing PLIP file, receptor-only input and a translated zero-contact control. GNINA and bust can exit zero for empty or bad ligand files while producing no valid result, so require file existence, parsed records and full coverage. A PLIP receptor-only XML has no binding site; it differs from a recognized, identity-matched ligand with zero contacts. The recorded calculation retained the original errors and corrected results; those records are not distributed with this article. GNINA emitted a space-groups.txt warning, and its SDF header caused RDKit to infer 3D from nonzero z coordinates; raw files and warnings were preserved, while all three structures passed finite-coordinate and chemical-identity checks.

5. Check every pose with PoseBusters

Linux

Copy the installed posebusters/config/redock.yml to $RUN_DIR/config/redock.yml and record its hash. Require one CSV row per SDF record and retain all booleans. Pose index 0 passed all 28 enabled PoseBusters checks. Indices 1 and 2 failed only rmsd_≤_2å; the other enabled checks were true. This distinction matters: their failure is a reference-pose criterion, not an assertion that all geometry is invalid. All three were sent to PLIP; indices 1 and 2 retain validation_failed. PLIP confirmed LIG:Z:1, the same E20 InChIKey and 28 heavy atoms for each. No pose or failure was silently excluded.

Do not overwrite the recorded run or environment; use fresh versioned paths.

PROJECT_DIR=/path/to/molecular-docking-validation
RUN_DIR="$PROJECT_DIR/runs/example-run"
INPUT_DIR="$PROJECT_DIR/data/prepared"
cd "$PROJECT_DIR"
mkdir -p "$RUN_DIR"/{dock,config,validation,complex,interactions}
bust "$RUN_DIR/dock/poses.sdf" -l "$INPUT_DIR/reference.sdf" \
  -p "$INPUT_DIR/receptor.pdb" --config "$RUN_DIR/config/redock.yml" \
  --max-workers 0 --outfmt csv --output "$RUN_DIR/validation/posebusters.csv"
Official source

Expected result

Three rows; all enabled checks pass for index 0

Actual negative runs covered missing/invalid receptor, empty/invalid SDF, missing PLIP file, receptor-only input and a translated zero-contact control. GNINA and bust can exit zero for empty or bad ligand files while producing no valid result, so require file existence, parsed records and full coverage. A PLIP receptor-only XML has no binding site; it differs from a recognized, identity-matched ligand with zero contacts. The recorded calculation retained the original errors and corrected results; those records are not distributed with this article. GNINA emitted a space-groups.txt warning, and its SDF header caused RDKit to infer 3D from nonzero z coordinates; raw files and warnings were preserved, while all three structures passed finite-coordinate and chemical-identity checks.

6. Assemble and inspect every pose with PLIP

Linux

Before invoking PLIP, combine each selected SDF pose with the prepared receptor in the same coordinate frame. Preserve ligand bond orders; assign ligand LIG:Z:1 and update atom serials, TER and CONECT consistently. Assembly scripts are not distributed here, so the PDB complexes are a prerequisite to the command. The displayed command analyzes pose index 0 as pose_001_complex.pdb. In the recorded calculation, indices 0, 1 and 2 were assembled and analyzed separately in complex/pose_XXX_complex.pdb and interactions/pose_XXX. Require each XML to identify LIG:Z:1, the E20 InChIKey and 28 heavy atoms; parse the XML and nonempty text, retaining validation_failed for indices 1 and 2. Actual report names are report.xml/report.txt.

Do not overwrite the recorded run or environment; use fresh versioned paths.

PROJECT_DIR=/path/to/molecular-docking-validation
RUN_DIR="$PROJECT_DIR/runs/example-run"
INPUT_DIR="$PROJECT_DIR/data/prepared"
cd "$PROJECT_DIR"
mkdir -p "$RUN_DIR"/{dock,config,validation,complex,interactions}
plip -f "$RUN_DIR/complex/pose_001_complex.pdb" -x -t \
  -o "$RUN_DIR/interactions/pose_001"
Official source

Expected result

Three identity-matched XML/TXT reports

Actual negative runs covered missing/invalid receptor, empty/invalid SDF, missing PLIP file, receptor-only input and a translated zero-contact control. GNINA and bust can exit zero for empty or bad ligand files while producing no valid result, so require file existence, parsed records and full coverage. A PLIP receptor-only XML has no binding site; it differs from a recognized, identity-matched ligand with zero contacts. The recorded calculation retained the original errors and corrected results; those records are not distributed with this article. GNINA emitted a space-groups.txt warning, and its SDF header caused RDKit to infer 3D from nonzero z coordinates; raw files and warnings were preserved, while all three structures passed finite-coordinate and chemical-identity checks.

7. Run failure cases and freeze the manifest

Linux

Actual negative runs covered missing/invalid receptor, empty/invalid SDF, missing PLIP file, receptor-only input and a translated zero-contact control. GNINA and bust can exit zero for empty or bad ligand files while producing no valid result, so require file existence, parsed records and full coverage. A PLIP receptor-only XML has no binding site; it differs from a recognized, identity-matched ligand with zero contacts. The recorded calculation retained the original errors and corrected results; those records are not distributed with this article. GNINA emitted a space-groups.txt warning, and its SDF header caused RDKit to infer 3D from nonzero z coordinates; raw files and warnings were preserved, while all three structures passed finite-coordinate and chemical-identity checks. For a new calculation, keep failure controls in separate output directories and compare the successful path, source identities, input/output hashes and library versions. No failure-control or manifest-generation script is distributed here. Successful computational execution does not establish biological activity; no scientific experiments were performed.

Do not overwrite the recorded run or environment; use fresh versioned paths.

Official source

Expected result

Expected negative cases rejected and manifest hashes match

Actual negative runs covered missing/invalid receptor, empty/invalid SDF, missing PLIP file, receptor-only input and a translated zero-contact control. GNINA and bust can exit zero for empty or bad ligand files while producing no valid result, so require file existence, parsed records and full coverage. A PLIP receptor-only XML has no binding site; it differs from a recognized, identity-matched ligand with zero contacts. The recorded calculation retained the original errors and corrected results; those records are not distributed with this article. GNINA emitted a space-groups.txt warning, and its SDF header caused RDKit to infer 3D from nonzero z coordinates; raw files and warnings were preserved, while all three structures passed finite-coordinate and chemical-identity checks.

Troubleshooting

Actual negative runs covered missing/invalid receptor, empty/invalid SDF, missing PLIP file, receptor-only input and a translated zero-contact control. GNINA and bust can exit zero for empty or bad ligand files while producing no valid result, so require file existence, parsed records and full coverage. A PLIP receptor-only XML has no binding site; it differs from a recognized, identity-matched ligand with zero contacts. The recorded calculation retained the original errors and corrected results; those records are not distributed with this article. GNINA emitted a space-groups.txt warning, and its SDF header caused RDKit to infer 3D from nonzero z coordinates; raw files and warnings were preserved, while all three structures passed finite-coordinate and chemical-identity checks.

Still not working

Alternatives

Uni-Dock has not been validated in this example; it requires independent checks of its inputs, GPU requirements, preparation and scoring. Reference self-checks cannot replace docking.

Related guides

Cite this page

ChemAI Atlas. Dock, Validate, and Inspect a Protein–Ligand Pose Locally. Updated . https://chemaiatlas.com/recipes/local-molecular-docking-pose-validation/