proteinmpnn

Name: proteinmpnn
Availability: InStock
Author: zongtingwei

Design protein sequences via ProteinMPNN inverse folding on given backbones.

25|5|Updated Mar 22, 2026

Search Tags:#bioinformatics #inverse-folding #sequence-design #protein-design #backbone-design #proteinmpnn

Ready to run

npx skills add https://github.com/zongtingwei/Bioclaw_Skills_Hub --skill proteinmpnn-zongtingwei

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: proteinmpnn
Source: https://github.com/zongtingwei/Bioclaw_Skills_Hub/tree/main/skills/protein-design/skills/proteinmpnn
Command: npx skills add https://github.com/zongtingwei/Bioclaw_Skills_Hub --skill proteinmpnn-zongtingwei

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Design protein sequences using ProteinMPNN inverse folding for efficient, targeted sequence design on given backbones, enabling redesigns, constraint-driven modifications, and multi-state optimization.

Core Features & Use Cases

Inverse folding-based sequence design with ProteinMPNN to generate high-quality variants for a specified backbone
Supports backbone design, targeted redesign, fixing specific residues, optimization for expression or stability, and multi-state/negative design
Interoperability with related design models (rfdiffusion/bindcraft for backbones, ligandmpnn for ligand-aware design, solublempnn for solubility tuning)
Biomodal workflow support via modal_ligandmpnn.py and typical inputs/outputs for downstream QC and design workflows

Quick Start

Design a set of protein sequences for a given backbone using ProteinMPNN with optional fixed residues and design constraints.

Frequently Asked Questions about proteinmpnn

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

FAQPage Schema

How do I design protein sequences for a specific backbone structure?▼

Protein sequence design for a specific backbone is performed using ProteinMPNN inverse folding, which generates high-quality sequences that fold into the given 3D structure. It supports targeted redesigns and constraint-driven modifications.

What is inverse folding and when do I need it for protein design?▼

Inverse folding is the process of predicting amino acid sequences that will fold into a desired backbone structure. You need it when designing novel proteins, redesigning existing sequences, or optimizing sequences for expression and stability.

Can I fix specific residues while redesigning the rest of a protein sequence?▼

Yes, you can fix specific residues while redesigning others using ProteinMPNN constraint-driven modifications. This allows targeted sequence optimization while maintaining critical structural or functional motifs.

Do I need a GPU environment to run ProteinMPNN for sequence design?▼

Yes, running ProteinMPNN requires Python 3.8+ and a compatible GPU environment with CUDA 11.x+ and sufficient VRAM. Access to backbone structures and design constraints is also necessary before execution.

Does ProteinMPNN support multi-state and negative design?▼

Yes, ProteinMPNN supports multi-state design and negative design to optimize sequences across multiple states or explicitly avoid undesired conformations. This enables advanced optimization for stability and expression.

What's the best way to optimize a protein sequence for stability and expression?▼

Optimizing a protein sequence for stability and expression is best achieved using ProteinMPNN inverse folding with multi-state design capabilities. It generates variants tailored to your backbone while applying specific optimization constraints.