> ## 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. # DBP Design Metrics > Score protein-DNA designs with dbp_design metrics and configurable failure behavior
DBP Design Metrics
**License:** This constraint can use multiple tools, each under its own license. See the **Tools Used** tab and each tool's page for license details.

This constraint 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.

evo-design/proto-language/proto\_language/constraint/protein\_structure/dbp\_design\_metrics\_constraint.py
View source
Constraint contributors
Evaluate dbp\_design metrics and return normalized penalties in `[0, 1]`. Resolves (reuses or predicts) one protein-operator complex PDB per candidate tuple, scores it with the heavy `dbp_design` PyRosetta metric suite, and converts the extracted metrics into a weighted-composite penalty where `0` is best and `1` is worst. When evaluation fails and `fail_hard` is disabled, returns `failure_score` and records the error in metadata. ## API Reference
ConfigDBPDesignMetricsConfig Source
Config for the dbp-design-metrics constraint. Predicts (or reuses) a protein-DNA complex and scores it with the heavy `dbp_design` metric suite (PyRosetta H-bond typing, compactness, interface ddG / CMS / shape complementarity / packstat / buried unsats / RotamerBoltzmann, and notebook-aligned charge ratios). The score is the weight-normalized sum of per-metric penalties clamped to `[0, 1]`, where `0` means every threshold is satisfied and `1` means none are. Path to local dbp\_design repository root; required (no default). Relative path to count\_hbond\_types.py script. Relative path to compactness\_filter.py script. Optional site-packages path containing pyrosetta when not in the active env. If true, raise on dbp metric tool failures instead of returning failure\_score. Penalty score returned when dbp metric evaluation fails in fail-soft mode. Path to calibrated ddG/CMS sigmoid prefilter equation file; None skips the prefilter. Path to prefilter log-prob cutoff file; used when prefilter\_cut is None. Explicit prefilter log-prob cutoff; when None it is loaded from prefilter\_cut\_path. HBond energy cutoff passed to count\_hbond\_types. Minimum weighted base score required. Minimum weighted phosphate score required. Minimum weighted bidentate score required. Minimum backbone-phosphate H-bond contacts. Minimum per-SSE contacts from compactness filter (default makes this penalty component inactive). Maximum allowed loop length from compactness filter (default makes this penalty component inactive). Require motif\_in\_rec\_helix metric to be true. Require rifres\_in\_rec\_helix metric to be true. Maximum allowed buried unsatisfied polar atoms at the interface. Minimum interface shape complementarity score (Sc). Minimum packstat score for foldability/packing quality. Minimum max RotamerBoltzmann-style score over interface RKQE residues. Minimum avg\_top\_two\_rboltz score. Maximum allowed interface ddG (more negative is better). Minimum required contact molecular surface (CMS). Lower bound of the allowed net\_charge\_over\_sasa window. Upper bound of the allowed net\_charge\_over\_sasa window. Maximum allowed ddg/contact\_molecular\_surface ratio. Normalized distance from threshold at which heavy penalties reach full penalty. Weight for base score penalty component. Weight for phosphate score penalty component. Weight for bidentate score penalty component. Weight for backbone phosphate contact penalty. Weight for compactness contact penalty. Weight for loop-length penalty component. Weight for motif-in-recognition-helix check. Weight for rifres-in-recognition-helix check. Weight for buried-unsatisfied-polars penalty. Weight for shape-complementarity penalty. Weight for packstat penalty. Weight for max-RKQE RotamerBoltzmann preorganization penalty. Weight for avg-top-two RotamerBoltzmann preorganization penalty. Weight for ddG notebook-alignment penalty. Weight for CMS notebook-alignment penalty. Weight for net\_charge\_over\_sasa notebook-alignment penalty. Weight for ddg\_over\_cms notebook-alignment penalty. Maximum allowed MPNN score when available in candidate metadata. Penalty used when MPNN score is unavailable in candidate metadata. Weight for MPNN score threshold penalty. Predictor for the protein-DNA complex; must be DNA-capable (alphafold3/boltz2/protenix). Options: `esmfold`, `esmfold2`, `alphafold3`, `boltz2`, `chai1`, `protenix`, `alphafold2`, `alphafold2_binder` Configuration for ESMFold structure prediction. Configuration for ESMFold2 structure prediction. Configuration for AlphaFold3 structure prediction. Configuration for Boltz2 structure prediction. Configuration for Chai1 structure prediction. Configuration for Protenix structure prediction. Configuration for the general AlphaFold2 multimer structure predictor. Configuration for the AF2 binder-design backend.
ReturnsConstraintOutput
Per-candidate score in `[0, 1]` (lower is better) with `dbp_metrics`, `dbp_penalties`, `pdb_path` (and `dbp_metrics_error` on soft failure) metadata.
## Usage ```python python icon="python" theme={null} from proto_language.core import Constraint from proto_language.constraint import dbp_design_metrics_constraint, DBPDesignMetricsConfig constraint = Constraint( inputs=[segment], function=dbp_design_metrics_constraint, function_config=DBPDesignMetricsConfig( # Configure parameters here ), ) scores = constraint.evaluate() ``` ## Metadata | Property | Value | | --------------- | ------------------------------- | | Key | `dbp-design-metrics` | | Function | `dbp_design_metrics_constraint` | | Category | `protein_structure` | | Mode | `discrete` | | Uses GPU | `True` | | Supported Types | `protein`, `dna` |