USRCAT (RDKit)
Compare molecules using USR/USRCAT shape fingerprints to shortlist candidates from a large library for further review.
API/CLI tool ID: usrcat
Understanding inputs and outputs
USRCAT summarizes molecular shape and distributions of atom types as a short list of numbers, called a shape fingerprint. It compares these summaries to quickly shortlist similar candidates from a large collection, without a detailed atom-by-atom alignment.
Place one reference molecule in querySdf and the candidate collection in libraryFiles. Both inputs require V2000 SDF files with actual three-dimensional coordinates. SDF stores atoms, bonds and coordinates; an .sdf extension alone does not mean that a 3D structure is ready.
topN is the number of top candidates returned across all input files combined. A usrcat_score closer to 1 indicates greater similarity in this fingerprint, not shared biological activity or binding ability. Its calculation differs from shape-it, so the numerical scores should not be compared directly.
Input settings
| Setting | Format | Required | Default | Description and constraints |
|---|---|---|---|---|
querySdf | sdf | Required | — | One reference molecule: a V2000 SDF with 3D coordinates — Extensions sdf |
libraryFiles | file | Required | — | 1-8 SDF files to search (V2000 with finite 3D coordinates, at most 10,000 molecules in total). A 3D flag does not guarantee a physically valid structure — Extensions sdf, JSON array |
topN | number | Optional | 50 | Number of top similar molecules to export (across all files) — Minimum 1, Maximum 10000 |
Output metrics
| Column | Display name | Range | Description |
|---|---|---|---|
usrcat_score | USRCAT similarity | 0~1 | Closer to 1 means more similar shape and atom-type distributions, not precise alignment, binding strength or biological activity. If atoms tie in distance when reference points are selected for the shape summary, changing their order in the file can change the score even with the same coordinates and bonds. |
See Reading results for metric definitions and interpretation. A range shows possible values; whether higher or lower is better depends on the metric.
Execution availability
Open an execution plan in your project to see available tool versions and costs. Review and approve the inputs, settings, and terms before running.
License
Code, model weights, and accompanying data have their own terms. The table summarizes terms from official sources; check your execution plan for the configuration you will use. License texts are available in Third-party notices.
| Component | License | Commercial use | Source | Notes |
|---|---|---|---|---|
| Code | BSD-3-Clause | Allowed, subject to notice requirements | License text | USR/USRCAT is included in RDKit and is covered by the RDKit license. |
Checked on 2026-10-26.
References
TNP
Calculate developability metrics from nanobody (VHH) sequences. Use them to compare candidate properties; they do not guarantee successful drug development.
Connecting AI tools and installing skills
Connect the Nadoo Bio MCP server to AI tools such as Claude Code, Codex, and Cursor, and install agent skills.