一键导入
boltz
Use when predicting biomolecular structures (proteins, RNA, DNA, ligands) with the open-source Boltz diffusion model as an alternative to AlphaFold3.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Use when predicting biomolecular structures (proteins, RNA, DNA, ligands) with the open-source Boltz diffusion model as an alternative to AlphaFold3.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Onboard and manage Paperclip AI for research-paper knowledge and agent orchestration
Generate a structured scientific post and publish it to Infinite. Runs a focused single-agent investigation (PubMed search → LLM analysis → hypothesis/method/findings/conclusion) and posts the result. Faster than scienceclaw-investigate — best for targeted, single-topic posts.
Infinite platform integration for AI agent collaboration
Read a CSV or XLSX file and return columns, shape, dtypes, and first N rows as JSON.
Execute arbitrary Python code and return stdout. NumPy, pandas, scipy, matplotlib, and other scientific libraries are available.
Generate a structured scientific PDF report from a JSON description. Accepts a JSON file specifying title, authors, abstract, sections (headings, text, tables, figures), and inline data panels (heatmap, bar, scatter, line). Produces a publication-style A4 PDF using reportlab with no LaTeX dependency. All figures are either loaded from PNG paths or generated on-the-fly from inline data.
| name | boltz |
| description | Use when predicting biomolecular structures (proteins, RNA, DNA, ligands) with the open-source Boltz diffusion model as an alternative to AlphaFold3. |
| metadata | null |
Predict biomolecular structures using Boltz, an open-source diffusion model. Boltz handles proteins, RNA, DNA, small molecules, ions, and covalent modifications in a single model without requiring multiple sequence alignments (MSA-optional). It serves as a strong open-source alternative to AlphaFold3.
pip install boltz
Boltz uses YAML for flexible entity specification:
# complex.yaml — protein + ligand
version: 1
sequences:
- protein:
id: A
sequence: MTEYKLVVVGAGGVGKSALTIQLIQNHFVDEYDPTIEDSY...
- ligand:
id: B
smiles: "CC1=CC=C(C=C1)S(=O)(=O)N" # or CCD code
ccd: ATP # alternative: use CCD code
# binder-target complex
version: 1
sequences:
- protein:
id: [A, B] # homodimer
sequence: MTEYKLVVVGAGGVGKS...
count: 2
- protein:
id: C
sequence: EVQLVESGGGLVQPGG... # binder
# Single prediction
boltz predict complex.yaml \
--out_dir results/ \
--accelerator gpu \
--devices 1 \
--num_workers 4
# Batch prediction (multiple YAML files)
boltz predict inputs/ \
--out_dir results/ \
--accelerator gpu
# Without MSA (faster, slightly lower accuracy for monomers)
boltz predict complex.yaml \
--out_dir results/ \
--use_msa_server false
from boltz.main import predict
predict(
data="complex.yaml",
out_dir="results/",
accelerator="gpu",
devices=1,
num_predictions=1, # ensemble size
recycling_steps=3,
diffusion_samples=1
)
results/
boltz_results_complex/
predictions/
complex/
complex_model_0.cif # Predicted structure (CIF format)
complex_confidence_model_0.json # Confidence scores
lightning_logs/ # Training logs (ignore)
import json
with open("complex_confidence_model_0.json") as f:
conf = json.load(f)
# Key metrics
plddt = conf["plddt"] # Per-residue confidence (0-100)
ptm = conf["ptm"] # Global fold confidence (0-1)
iptm = conf["iptm"] # Interface confidence (0-1)
ligand_iptm = conf.get("ligand_iptm") # Ligand interface confidence
pde = conf.get("pde") # Predicted Distance Error
print(f"pTM={ptm:.3f}, ipTM={iptm:.3f}")
| Metric | Marginal | Acceptable | Good |
|---|---|---|---|
| pLDDT (mean) | <60 | 60–80 | >80 |
| ipTM | <0.5 | 0.5–0.7 | >0.7 |
| pTM | <0.4 | 0.4–0.6 | >0.6 |
| Feature | Boltz | AF2 | AF3 |
|---|---|---|---|
| Open source | ✓ | ✓ (weights) | ✗ |
| Ligands | ✓ | ✗ | ✓ |
| RNA/DNA | ✓ | ✗ | ✓ |
| MSA required | Optional | Yes | Optional |
| Local run | ✓ | ✓ | Limited |
| CIF output | ✓ | PDB | CIF |
# Using BioPython
python3 -c "
from Bio.PDB import MMCIFParser, PDBIO
parser = MMCIFParser()
structure = parser.get_structure('pred', 'complex_model_0.cif')
io = PDBIO()
io.set_structure(structure)
io.save('complex_model_0.pdb')
"