> ## Documentation Index > Fetch the complete documentation index at: https://proto.evodesign.org/docs/llms.txt > Use this file to discover all available pages before exploring further. # DSSP > [DSSP](https://github.com/PDB-REDO/dssp) (Dictionary of Protein Secondary Structure) is a standard program for assigning protein secondary structure from atomic coordinates. It identifies hydrogen-bonding and geometric patterns in a protein backbone and labels each residue as helix, strand, turn, or another defined state. This toolkit runs DSSP and collapses those assignments into helix, sheet, and loop percentages for a selected chain in each input structure.
DSSP
**License:** DSSP is open source and free for academic and commercial use under a BSD-2-Clause license. Please refer to [the license](https://github.com/PDB-REDO/dssp/blob/trunk/LICENSE) for full terms.

This toolkit is open source. Any third-party models, product names, or trademarks referenced are the property of their respective owners, and Proto is not affiliated with them.

PDB-REDO/dssp
PDB-REDO/dssp
View repo
DSSP 4: FAIR annotation of protein secondary structure
Maarten L. Hekkelman, Daniel Alvarez Salmoral, ... Robbie P. Joosten
Protein Science (2025)
Read paper
```bibtex theme={null} @article{hekkelman_2025_dssp4, title={{DSSP} 4: FAIR annotation of protein secondary structure}, author={Hekkelman, Maarten L. and {\'A}lvarez Salmoral, Daniel and Perrakis, Anastassis and Joosten, Robbie P.}, journal={Protein Science}, volume={34}, number={8}, pages={e70208}, year={2025}, doi={10.1002/pro.70208}, pmid={40671631}, pmcid={PMC12268231}, } @article{kabsch_1983_dssp, title={Dictionary of protein secondary structure: Pattern recognition of hydrogen-bonded and geometrical features}, author={Kabsch, Wolfgang and Sander, Christian}, journal={Biopolymers}, volume={22}, number={12}, pages={2577--2637}, year={1983}, doi={10.1002/bip.360221211}, pmid={6667333}, } ```
Copy citation
evo-design/proto-tools/proto\_tools/tools/structure\_scoring/dssp
View source
Open Notebook
Open notebook
DSSP Secondary Structure
Toolkit contributors
| Function | Description | | | -------------------------------- | --------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `run_dssp_secondary_structure()` | Assign helix/sheet/loop percentages using the DSSP binary | Docs Source | ## Background DSSP ([Kabsch and Sander, 1983](https://doi.org/10.1002/bip.360221211)) is an assignment program that classifies each residue of a protein into a secondary-structure state by inspecting the geometry and the hydrogen-bond pattern of the protein backbone. Recurring turns are assigned as helices (states `H`, `G`, and `I` for alpha, 3-10, and pi helices respectively), recurring bridges between residues form ladders that are assigned as strand (`E`), and isolated bridges, turns, bends, and unassigned residues form the remaining states. DSSP works only from atomic coordinates and does not predict secondary structure from sequence alone. The modern implementation ([Hekkelman et al., 2025](https://doi.org/10.1002/pro.70208)) is maintained by the [PDB-REDO project](https://github.com/PDB-REDO) at the Netherlands Cancer Institute and ships as the `mkdssp` command-line program with extended mmCIF support and FAIR annotation. This toolkit collapses the per-residue DSSP states into a coarse three-class summary for the chain of interest. Helix percentage counts residues assigned `H`, `G`, or `I`. Sheet percentage counts residues assigned `E`. Loop percentage counts every other DSSP state (`B`, `T`, `S`, the unassigned state, and the `P` polyproline-II state introduced in DSSP 4). The three percentages sum to 100 for the counted residues of the selected chain. ### Learning Resources * [PDB-REDO/dssp](https://github.com/PDB-REDO/dssp) (PDB-REDO project, Netherlands Cancer Institute). Official repository and the source of the `mkdssp` command-line program that this toolkit invokes. * [`mkdssp` command-line reference](https://github.com/PDB-REDO/dssp/blob/trunk/doc/mkdssp.md) (PDB-REDO project). Reference documentation for the DSSP state alphabet and the command-line interface of the program that this toolkit invokes. ## Tools
### DSSP Secondary Structure (`dssp-secondary-structure`) Assigns secondary structure with the `mkdssp` program for a selected chain in each input structure and returns the resulting helix, sheet, and loop percentages. Inputs are supplied as one or more `DSSPStructureInput` objects, each carrying a `Structure` (or a path / coordinate string accepted by `Structure`) plus the chain identifier to analyse. Multiple structures in one call are processed independently and the results are returned in input order. #### API Reference
Source Structures and chains to analyze. Chain to analyze. `None` analyzes the first chain in the structure. `mkdssp` always runs on the whole structure, so this only selects which chain's percentages are reported. Protein structure to analyze.
Source Verbosity level (0=quiet, 1=info, 2=debug, 3=raw subprocess stderr). `True` is coerced to `1` and `False` to `0`. Device to run the tool on. Maximum execution time in seconds. `None` waits indefinitely. Random seed. When set, tools run reproducibly up to small GPU float noise (see `BaseToolOutput.approx_equal`), and the seed participates in cache keys. When None, cacheable seed-sensitive tools skip cache until seeded.
Source Per-input secondary-structure percentages. Analyzed chain label in the input structure namespace. Name of the metric that best summarizes the result overall (e.g. `"avg_plddt"` for AlphaFold2). Used by downstream UI and reporting to pick a headline value. Concrete Metrics subclass tag; enables typed reconstruction after a serialization round-trip. **Metrics** (one set per `results` item) | Metric | Type | Range | Availability | | ----------- | ----- | ------------ | ------------ | | `helix_pct` | float | 0.0 to 100.0 | always | | `sheet_pct` | float | 0.0 to 100.0 | always | | `loop_pct` | float | 0.0 to 100.0 | always |
#### Applications This tool is appropriate for filtering designed proteins by their secondary-structure composition, for summarising the helical or beta-sheet content of a predicted structure ensemble, and for any analysis that requires a DSSP-backed secondary-structure assignment as an upstream step in a larger pipeline. The collapsed three-class summary is the right shape for ranking or filtering large structure batches by composition. #### Usage Tips * **Select the chain to analyze, or leave `chain` empty to analyze the first chain.** `chain` is a `SingleChainSelection` (e.g. `"A"`); each input structure yields one result row for its selected chain. The input validator hard-errors when the selected chain is not present and lists the available chains. Omitting `chain` analyzes the first chain in the structure. * **Structures with more than 62 chains are rejected.** The DSSP standalone runs on PDB-format text, which represents a chain identifier as a single character from `A-Z`, `a-z`, or `0-9`. Structures exceeding this limit cannot be dispatched through the wrapper. * **Helix percentage counts DSSP states `H`, `G`, and `I`.** Sheet percentage counts state `E`. Loop percentage counts every remaining state, including `B`, `T`, `S`, the unassigned state, and the `P` polyproline-II state introduced in DSSP 4. The three percentages sum to 100 for the counted residues of the selected chain. * **The first model is used for multi-model structures.** Only the first model parsed by Biopython contributes to the residue counts. To analyse a specific model in an NMR ensemble or a multi-state file, extract that model into its own `Structure` before passing it in.
## Toolkit Notes These apply to every DSSP tool in this toolkit (`dssp-secondary-structure`). * **Structure inputs accept either typed `Structure` objects or a path / coordinate string.** A field validator normalises raw paths and `Structure`-coercible values into `Structure` instances at input time. The wrapper writes each parsed structure to a temporary PDB file for the DSSP program and removes it after the call. * **Outputs are returned as typed metric objects.** Each `DSSPSecondaryStructureMetrics` result carries the analysed chain identifier and the three secondary-structure percentages, with `helix_pct`, `sheet_pct`, and `loop_pct` constrained to the range 0 to 100. Results serialise to CSV or JSON through the standard export interface. **Example notebook:** See the [full working example](https://github.com/evo-design/proto-tools/blob/47e34afa5ea240a3b406e323dc38aa5dc85f223e/proto_tools/tools/structure_scoring/dssp/examples/example.ipynb) for a copy-paste-ready walkthrough. ## Infrastructure Guides The following guides cover how to run tools efficiently and at scale. Keep a tool's model warm across calls instead of reloading it every invocation. How GPUs are allocated to tools and how to target specific devices. Fan a batch of inputs out across multiple GPUs.