Can I perform inverse folding with this skill?

Yes, the ESM skill supports inverse folding by providing a structure to the ESM3 model to generate corresponding viable protein sequences.

What is the difference between ESM3 and ESM C in this skill?

ESM3 is a multimodal generative model used for designing sequences and structures, while ESM C is optimized for creating efficient, high-quality protein embeddings for analysis.

Does this skill require a local GPU?

While local model usage (like esm3-sm-open-v1) performs best on a GPU, the skill also supports the cloud-based Forge API for scalable inference without local hardware requirements.

How do I get an API token for the cloud features?

To use the Forge API for larger models and batch processing, you can obtain a token at the EvolutionaryScale Forge platform website.

ESM Protein Design & Modeling

Name: ESM Protein Design & Modeling
Author: Sologa

bySologa

0•

Data Science & ML

Designs, analyzes, and generates protein sequences and structures using Evolutionary Scale Modeling (ESM3 and ESM C).

The ESM skill integrates state-of-the-art protein language models into Claude's workflow, enabling advanced biological research and protein engineering. It provides comprehensive support for ESM3, a generative multimodal model capable of co-designing protein sequence, structure, and function, as well as ESM C for efficient representation learning and embeddings. Whether you are performing inverse folding, predicting structures, or generating novel variants for drug discovery, this skill facilitates seamless interaction with both local models and the scalable Forge API for high-throughput inference.

Key Features

010 GitHub stars

02Automated structure prediction and inverse folding (sequence design from structure)

03High-performance protein embeddings and representations using ESM C models

04Cloud-based scalability via Forge API integration for large-scale processing

05Multimodal generative protein design across sequence, structure, and function tracks

06Function-conditioned generation for proteins with specific biological properties

Use Cases

01Designing novel functional proteins and enzymes with specific structural constraints

02Accelerating protein engineering through iterative chain-of-thought refinement

03Extracting biological features for downstream machine learning and function prediction

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add sologa/codex-pipeline esm

For use in Claude.ai and ChatGPT

Download Skill