Structural TM-score Similarity

Compare structure TM-score against a target. Returns 1 - TMscore.

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.

---

Tools Used Tools Used Source Source

Structure Prediction · 7 tools

proto-bio/proto-language/proto_language/constraint/protein_structure/structure_similarity_constraint.py

View source

Predicts structure and compares TM-score. Returns (1.0 - TMscore).

This constraint automatically selects the appropriate alignment tool based on the oligomer state of the inputs:

• Monomer vs monomer comparisons use standard TMalign.
• Comparisons involving multiple chains use USalign with -mm 1 and default values for all other parameters.

The two raw TM-scores are combined per config.tm_score_normalization (default: arithmetic mean of the proposal- and target-normalized scores); the target is the reference structure. Metadata describes the predicted full input tuple/complex, not an individual chain.

API Reference

ConfigStructureTMScoreConfig Source

Configuration for TM-score based structure similarity.

This configuration extends StructureSimilarityConfig for calculating the Template Modeling score (TM-score) between the target and proposal structures. TM-score is a metric for assessing the topological similarity of protein structures and is less sensitive to local variations than RMSD.

The constraint returns a score calculated as (1.0 - TM_score), where 0.0 indicates a perfect match (TM-score = 1.0) and values closer to 1.0 indicate poor structural similarity.

Inherits target specification (target_chains, target_structure, min_target_plddt) from StructureSimilarityConfig and tool selection (structure_tool, esmfold_config, esmfold2_config, alphafold3_config, boltz2_config, chai1_config, protenix_config, alphafold2_config, alphafold2_binder_config) from StructureBasedConstraintConfig.

plddt_threshold · number

Drop residues with pLDDT (0-100 scale, in B-factor) below this before alignment; None keeps all.

tm_score_normalization · enum · default: "mean"

How to combine the two TM-scores from TM-align/US-align: structure1, structure2, max, min, or mean.

Options: structure1, structure2, max, min, mean

structure_tool · enum · default: "esmfold"

Predictor: esmfold/esmfold2/alphafold3/boltz2/chai1/protenix/alphafold2/alphafold2_binder.

Options: esmfold, esmfold2, alphafold3, boltz2, chai1, protenix, alphafold2, alphafold2_binder

esmfold_config · ESMFoldConfig

Configuration for ESMFold structure prediction.

esmfold2_config · ESMFold2Config

Configuration for ESMFold2 structure prediction.

alphafold3_config · AlphaFold3Config

Configuration for AlphaFold3 structure prediction.

boltz2_config · Boltz2Config

Configuration for Boltz2 structure prediction.

chai1_config · Chai1Config

Configuration for Chai1 structure prediction.

protenix_config · ProtenixConfig

Configuration for Protenix structure prediction.

alphafold2_config · AlphaFold2Config

Configuration for the general AlphaFold2 multimer structure predictor.

alphafold2_binder_config · AlphaFold2BinderStructureConfig

Configuration for the AF2 binder-design backend.

target_chains · array | Complex

Target chains: a tuple of sequence strings (entity types auto-detected).

target_structure · Structure | string

Target structure: a Structure object, file path (.pdb/.cif), or raw PDB/CIF content string.

min_target_plddt · number · default: "0.6"

Min mean pLDDT (0-1 scale) for a target folded from sequence; ignored when target_structure is set.

Usage

python

from proto_language.core import Constraint
from proto_language.constraint import structure_tmscore_constraint, StructureTMScoreConfig

constraint = Constraint(
    inputs=[segment],
    function=structure_tmscore_constraint,
    function_config=StructureTMScoreConfig(
        # Configure parameters here
    ),
)

scores = constraint.evaluate()

Metadata

Property	Value
Key	structure-tmscore
Function	structure_tmscore_constraint
Category	protein_structure
Mode	discrete
Uses GPU	True
Supported Types	protein, rna, dna, ligand