boltz-structure-prediction

Predict protein and complex structures with Boltz-1 / Boltz-2 and output CIF files.

25|5|Updated Mar 22, 2026
One-click install
npx skills add https://github.com/zongtingwei/Bioclaw_Skills_Hub --skill boltz-structure-prediction
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: boltz-structure-prediction
Source: https://github.com/zongtingwei/Bioclaw_Skills_Hub/tree/main/skills/protein-design/skills/boltz-structure-prediction
Command: npx skills add https://github.com/zongtingwei/Bioclaw_Skills_Hub --skill boltz-structure-prediction

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Boltz-1 / Boltz-2 structure prediction for proteins, complexes, and ligand-aware validation.

Core Features & Use Cases

  • Open-source structure predictor for protein–protein and protein–ligand predictions.
  • Provides confidence metrics (pLDDT, pTM, ipTM) and CIF outputs for downstream QC and triage.
  • Use Case: validate designed protein interactions or compare alternatives against Alphafold2 Multimer or Chai predictions.

Quick Start

Run the Boltz structure prediction workflow with Modal or local install to generate predicted models and confidence metrics.

Frequently Asked Questions about boltz-structure-prediction

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I predict protein-ligand complex structures using open-source weights?

Predict protein-ligand complex structures using Boltz-1 or Boltz-2 open-source weights to generate CIF outputs with confidence metrics like pLDDT, pTM, and ipTM.

What GPU resources do I need for protein structure prediction with Boltz?

Protein structure prediction with Boltz requires Python 3.10+, CUDA 12.x, and a local GPU with at least 24GB VRAM, or you can run the workflow via Modal execution.

Can I validate designed protein interactions using Boltz structure prediction?

Validate designed protein interactions using Boltz structure prediction to generate confidence metrics and CIF outputs for downstream quality control and triage.

How does Boltz compare to Alphafold2 Multimer for protein-protein complex prediction?

Compare Boltz protein-protein complex predictions against Alphafold2 Multimer or Chai predictions to evaluate alternative structural models and validate designed interactions.

What output formats and confidence metrics does Boltz structure prediction produce?

Boltz structure prediction produces CIF output files containing predicted 3D models alongside confidence metrics including pLDDT, pTM, and ipTM scores for downstream analysis.

Are there limitations when running Boltz for protein structure prediction locally?

Local Boltz protein structure prediction is constrained by hardware requirements needing 24GB VRAM and CUDA 12.x, but Modal execution offers an alternative for running the workflow.