DesignCPU3 credits

ThermoMPNN — Stability Prediction

Predict how mutations affect protein thermostability

ThermoMPNN predicts the change in thermodynamic stability (DDG) caused by point mutations using transfer learning on ProteinMPNN features. Quickly classify mutations as stabilizing or destabilizing to guide protein engineering decisions.

$0.03

per API call

credits per call

/v1/design/stability

API endpoint

Features

DDG prediction for point mutations

Stabilizing vs destabilizing classification

Single and double mutant support

No GPU required — fast CPU inference

Sub-second response times

Transfer learning from ProteinMPNN features

Quick Start

ThermoMPNN — Python Examplepython

import requests

API_KEY = "sk-sci-your-key-here"
url = "https://scirouter.ai/v1/design/stability"

response = requests.post(url, json={
    "pdb": open("protein.pdb").read(),
    "mutations": ["A45G", "L120V", "S78T"]
}, headers={"Authorization": f"Bearer {API_KEY}"})

result = response.json()
for mut in result["predictions"]:
    label = "stabilizing" if mut["ddg"] < 0 else "destabilizing"
    print(f"{mut['mutation']}: DDG={mut['ddg']:.2f} kcal/mol ({label})")