What is the difference between ESM3 and ESM C?

ESM3 is a generative multimodal model used for designing sequences, structures, and functions, whereas ESM C is optimized for creating efficient protein embeddings and representations.

Does this skill support inverse folding?

Yes, it allows you to input a target 3D structure (via PDB) and generate sequences that are predicted to fold into that specific structure.

What is the Forge API in the context of ESM?

The Forge API is EvolutionaryScale's cloud platform that provides access to larger, high-performance models like ESM3-Large for scalable and high-throughput inference.

How do I install the required ESM dependencies?

You can install the core library using 'uv pip install esm'. For optimized local performance, installing 'flash-attn' is also recommended.

Can I use this skill without a cloud API?

Yes, the ESM skill supports local model usage, such as the open-weights esm3-sm-open-v1 model, for experimentation and local development.

Evolutionary Scale Modeling (ESM)

Name: Evolutionary Scale Modeling (ESM)
Author: henriquescastilho

byhenriquescastilho

•

数据科学与机器学习

Enables state-of-the-art protein language modeling for sequence design, structure prediction, and functional annotation.

The ESM skill integrates Evolutionary Scale Modeling (ESM3 and ESM C) into Claude, providing a comprehensive toolkit for computational biology and protein engineering. It allows researchers to design novel protein sequences with specific functional properties, predict 3D structures from sequences, perform inverse folding, and generate high-quality embeddings for machine learning tasks. Whether using local models for experimentation or the cloud-based Forge API for scalable inference, this skill streamlines the workflow for understanding and creating complex biological molecules within a developer environment.

主要功能

01High-accuracy protein structure prediction and inverse folding (sequence-from-structure)

02Function-conditioned generation and automated protein annotation

03Efficient protein representation learning using ESM C embeddings

041 GitHub stars

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

06Chain-of-thought refinement for iterative and complex protein engineering

使用场景

01Designing novel enzymes or fluorescent proteins with specific biological functions

02Extracting protein embeddings for functional classification and similarity analysis

03Predicting 3D structures and optimizing sequences for target structural folds

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add henriquescastilho/my-claude esm

For use in Claude.ai and ChatGPT