Run the bundled examples

Examples let you run Ultidock before supplying your own molecules. They use bundled receptor and prepared ligand PDBQT files; dataset examples also include dataset.json metadata describing their inputs. This is separate from docking/ligands.wget, the optional download list for your own runs. Example scripts pass --skip-wget so that download list cannot add molecules to the example.

Try one complete run

From an installed checkout with the required tools available:

ultidock example list
ultidock example run d2-antipsychotics --dry-run
ultidock example run d2-antipsychotics --mode auto

example list shows the runnable examples. --dry-run previews the selected files without docking. The real command creates a new examples/d2-antipsychotics/workspace/<timestamp>/ folder, copies the declared inputs, prepares the receptor, finds sites, builds grids and docks the ligands. It prints the workspace path. Existing example runs remain available for inspection; installed native tools are reused.

Open RESULTS_DIR/ inside that workspace for the SQLite score database and CSV tables. The corresponding 3D output poses are in DOCKING_DIR/. Follow First Docking Run to inspect a pose and Results and reports for the output files.

Choose another example

Example	What it demonstrates	Command
quickstart	Interactive guided D2 run	ultidock example run quickstart
d2-antipsychotics	Dopamine D2 receptor with three comparison ligands	ultidock example run d2-antipsychotics --mode auto
sert-escitalopram	Serotonin transporter with one ligand	ultidock example run sert-escitalopram --mode auto
gabaa-benzos	GABA-A receptor and ligand workflow	ultidock example run gabaa-benzos --mode auto
gabaa-8dd2-cav-emps	Site prediction against withheld known sites	ultidock example run gabaa-8dd2-cav-emps --dry-run

The D2 example stages 6CM4-edited.pdbqt, haloperidol, escitalopram and morphine. Its dataset notes give input provenance; the comparison molecules are not validated decoys. The GABA 8DD2 study withholds five GABA/zolpidem sites and compares controlled CaV-EMPS settings with fpocket and P2Rank. Read its example README before interpreting the results.

Bundled examples have different defaults: D2 and the interactive teacher start in CPU mode, while SERT and GABA-A require a GPU by default. The --mode auto commands above select a visible compatible GPU for AutoDock-GPU or fall back to CPU Vina; the teacher accepts only cpu or gpu. On a CUDA machine with multiple NVIDIA GPUs, docking jobs are spread across detected devices. Pass --mode gpu when GPU execution is required.

The quickstart teacher explains each stage before launching a real D2 run. Press Enter to advance, b to go back and q to exit. Preview all lessons with ultidock example run quickstart --dry-run. Use --yes only for explicit unattended execution. See Quick Start.

Try another site finder

D2 and SERT use Ultidock’s included CaV-EMPS finder by default. Pass fpocket or p2rank after the example name to compare a different prediction method:

ultidock example run sert-escitalopram fpocket --mode auto
ultidock example run sert-escitalopram p2rank --mode auto

These methods need their optional local tools; see Installation. P2Rank 2.5 needs Java 17–23; the documented Ubuntu choice is openjdk-21-jre-headless. If Java 25 is selected, it may fail with Unsupported class file major version 69. The bundled external/bin/prank launcher uses /usr/lib/jvm/java-21-openjdk-amd64 when JAVA_HOME is unset; set JAVA_HOME to Java 21 if it points to an incompatible installation.

The predictor writes new site boxes before docking. Each site needs its own AutoGrid map set; [autogrid] prepared N site grids means grid preparation finished. Docking work grows with the number of ligands times the number of sites. Keep the chosen method, tool version and raw prediction output with comparisons. The example scripts can also serve as automation templates.