> ## 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. # AF3 Off-target ipTM Specificity > Prefer high target ipTM and lower off-target ipTM using AF3
AF3 Off-target ipTM Specificity
**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/af3\_offtarget\_iptm\_specificity\_constraint.py
View source
Constraint contributors
Score AF3-based off-target specificity margin using ipTM. For each candidate tuple, builds the protein + target-DNA complex and one protein + off-target-DNA complex per off-target motif (motif substituted at `dna_indices`), folds them all in a single `predict_structures` batch, and reads `iptm` from each. The score is the specificity margin `target_iptm - best_off_target_iptm` against `desired_margin`: 0.0 when the margin is met or exceeded, 1.0 when there is no advantage. **Supported tools**: AlphaFold3 (any ipTM-producing structure tool configured via `structure_tool`). ## API Reference
ConfigAF3OffTargetIPTMSpecificityConfig Source
Configuration for the AF3 off-target ipTM specificity constraint. Full forward DNA scaffold (A/C/G/T) used for target folding. Target motif in A/C/G/T placed at dna\_indices in the scaffold. Off-target motifs substituted at dna\_indices for AF3 specificity. 0-based indices of motif positions in target\_dna\_sequence. Desired margin target\_iptm - best\_off\_target\_iptm. Include reverse-complement DNA strand for dsDNA AF3 folding. 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 specificity score in `[0, 1]` (lower is better) with `target_iptm` / `best_off_target_iptm` / `iptm_advantage` / `desired_margin` metadata (plus an `iptm_error` flag when the predictor omits `iptm`). The predicted target complex attaches to slot 0.
## Usage Off-target specificity for a designed DNA-binding protein with AF3: ```python python icon="python" theme={null} >>> from proto_language.core import Segment >>> binder = Segment(length=120, sequence_type="protein") >>> specificity = Constraint( ... inputs=[binder], ... function=af3_offtarget_iptm_specificity_constraint, ... function_config={ ... "structure_tool": "alphafold3", ... "target_dna_sequence": "ACGTACGTACGT", ... "target_motif": "GTAC", ... "off_target_motifs": ["AAAA", "TTTT"], ... "dna_indices": [2, 3, 4, 5], ... }, ... ) ``` ## Metadata | Property | Value | | --------------- | ------------------------------------------- | | Key | `af3-offtarget-iptm-specificity` | | Function | `af3_offtarget_iptm_specificity_constraint` | | Category | `protein_structure` | | Mode | `discrete` | | Uses GPU | `True` | | Supported Types | `protein`, `dna` |