MolScope Toolkit Open-source medchem property tables • reports • picklists

MolScope Toolkit: from SMILES to practical medchem triage.

Created by Elokely Lab • University of Wyoming School of Pharmacy

MolScope Toolkit turns SMILES lists into analysis-ready CSV tables and then produces category-focused summaries for CNS MPO, oral filters, med‑chem alerts, toxicity flags, metabolism soft spots, PK heuristics, developability indices, optional 3D descriptors, series analytics, and more.

CI status Issues Discussions

Install

For the calculator features you’ll need RDKit. The most reliable public install path is conda‑forge (or Docker). If you already have a compatible results CSV from another pipeline, you can run analyzers without RDKit.

Recommended: conda (RDKit‑friendly)

git clone https://github.com/kelokely/-molprop-toolkit.git molprop-toolkit
cd molprop-toolkit

conda env create -f environment.yml
conda activate molprop-toolkit

# editable install with console scripts + Parquet support
pip install -e ".[dev,parquet]"

Quick run (example)

# calculator → results table (CSV or Parquet)
molprop-calc-v5 examples/example.smi -o results.parquet

# summary report (md + html)
molprop-report results.parquet

# picklists
molprop-picklists results.parquet --html

CLI tools

Full tool reference: tools/

The toolkit ships multiple console commands. Most workflows start with a calculator (v4 or v5), then use analyzers and report builders to generate shareable outputs.

Calculators

Generate a wide results table from SMILES.

  • molprop-calc-v4 – baseline medchem table
  • molprop-calc-v5 – adds solubility/permeability/PK + optional 3D
  • Ligand preparation guide – recommended prep sequences and which SMILES to use downstream

Analyze + report

Summaries by category and shareable reports.

  • molprop-analyze – master analyzer
  • molprop-report – Markdown/HTML report bundle
  • molprop-sketch – 2D depictions + HTML index

Operational tools

Series analytics and decision picklists.

  • molprop-series – scaffolds + clustering
  • molprop-picklists – configurable picklists + HTML
  • molprop-similarity – fingerprint-based similarity search + diversity picking
  • molprop-featurize – ML-ready feature packs (fingerprints, fragments, FG, DL exports)
  • molprop-retro – offline retrosynthesis wrapper (AiZynthFinder)

Documentation

The repo includes detailed column references and narrative descriptions of analysis categories. The links below open the canonical sources on GitHub.

README

Installation, quickstart, and workflows.

Columns reference

Generated schema-driven column documentation.

Feature families

Prefix-based “feature-to-columns” cheat sheet.

Category narratives

Interpretation notes for major analysis categories.

Contributing

Contributions are welcome. If you have an idea for a new property family, analyzer, picklist, or report section, open an issue with an example CSV and a short description of the decision you want the tool to support.

Open an issue View pull requests