cmxflow 🧪
Build cheminformatics and computational chemistry pipelines with composable blocks. Tune end-to-end with Bayesian Optimization. Or ask an LLM agent to do it.
Quick examples
Prepare ligands for docking
from cmxflow import Workflow
from cmxflow.sources import MoleculeSourceBlock
from cmxflow.operators import (
MoleculeStandardizeBlock,
IonizeMoleculeBlock,
EnumerateStereoBlock,
ConformerGenerationBlock,
)
from cmxflow.sinks import MoleculeSinkBlock
workflow = Workflow()
workflow.add(
MoleculeSourceBlock(),
MoleculeStandardizeBlock(),
IonizeMoleculeBlock(),
EnumerateStereoBlock(),
ConformerGenerationBlock(),
MoleculeSinkBlock(),
)
workflow("library.smi", "prepared.sdf")
Dock a congeneric series
Pure-Python docking. Free docking is the default (index_poses=False); scaffold-indexed mode caches poses by Bemis–Murcko scaffold for ~3× faster throughput on congeneric series with consistent pose alignment.
from cmxflow import Workflow
from cmxflow.sources import MoleculeSourceBlock
from cmxflow.operators import ConformerGenerationBlock, MoleculeDockBlock
from cmxflow.sinks import MoleculeSinkBlock
from cmxflow.utils.parallel import make_parallel
workflow = Workflow()
workflow.add(
MoleculeSourceBlock(),
ConformerGenerationBlock(),
make_parallel(
MoleculeDockBlock(
receptor="receptor.pdb",
site_reference="crystal_ligand.sdf",
index_poses=True,
)
),
MoleculeSinkBlock(),
)
workflow("library.smi", "docked.sdf")
Tune a ligand-based virtual screen
from cmxflow import Workflow
from cmxflow.sources import MoleculeSourceBlock
from cmxflow.operators import MoleculeSimilarityBlock
from cmxflow.scores import EnrichmentScoreBlock
from cmxflow.opt import Optimizer
workflow = Workflow()
workflow.add(
MoleculeSourceBlock(),
MoleculeSimilarityBlock(queries="crystal_ligand.sdf"),
EnrichmentScoreBlock(target="active"),
)
opt = Optimizer(workflow, "benchmark.csv")
opt.optimize(n_trials=30, direction="maximize")
print(f"Best enrichment AUC: {opt.best_score:.3f}")
print(opt.best_params)
The four fingerprint parameters above are searched automatically — every block exposes its mutable parameters to the optimizer.
Or build it conversationally via an LLM agent
claude mcp add cmxflow -- cmxflow-mcp
"How many of the molecules in library.csv pass Lipinski's rules?"
"I need to build a ligand-based virtual screening workflow. I'm not sure if 2D or 3D is better. Can you optimize two workflows?"
"Dock the molecules in hits.csv against receptor.pdb with crystal_ligand.sdf as a reference."
The agent can build, run, and optimize workflows. See Using with Claude for full transcripts.
What's in the box
- 15+ blocks for sourcing, transforming, filtering, clustering, scoring, and docking molecules
- Bayesian optimization of pipeline parameters via Optuna
- Parallel execution for compute-heavy blocks (conformer generation, docking)
- Workflow serialization for save / load / reuse
- An MCP server with five tools:
build_workflow, run_workflow, optimize_workflow, manage_workflows, view_structures
Install
MCP server
claude mcp add cmxflow -- cmxflow-mcp
Optional: PyMOL
Required only for the view_structures MCP tool (3D visualization):
conda install -c conda-forge pymol-open-source
Documentation
Project
MIT licensed. See CONTRIBUTING.md and RELEASING.md.