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

ESM3 is a generative multimodal model used for design across sequence, structure, and function, while ESM C is an encoder-only model optimized for efficient protein embeddings and representations.

How does function-conditioned generation work?

You can provide functional annotations or labels as a prompt, and ESM3 will generate protein sequences that are predicted to possess those specific biological functions.

Does this skill require a paid API key?

The skill supports both local model usage (like esm3-sm-open-v1) and the cloud-based Forge API. While local models are free to run, the Forge API requires a token from EvolutionaryScale for high-performance inference.

Is this skill suitable for high-throughput screening?

Yes, it includes batch processing capabilities via the Forge API's async executor, allowing for the efficient processing of multiple protein sequences simultaneously.

Can I predict a protein's 3D structure from its sequence?

Yes, the ESM3 structure track allows for structure prediction from sequence, as well as inverse folding, which generates a protein sequence based on a target 3D structure.

Evolutionary Scale Modeling (ESM)

Name: Evolutionary Scale Modeling (ESM)
Author: workonthestreets

byworkonthestreets

0•

데이터 과학 및 ML

Accelerates protein engineering and design using ESM3 and ESM C models for sequence generation, structure prediction, and representation learning.

The ESM skill provides a comprehensive toolkit for protein language modeling, enabling developers and researchers to perform state-of-the-art biological design directly within their workflow. It leverages the ESM3 family for multimodal generative tasks—including sequence completion, 3D structure prediction, and inverse folding—alongside ESM C for generating high-quality protein embeddings for downstream machine learning. Whether performing local inference with open weights or utilizing the scalable Forge API for massive throughput, this skill streamlines complex protein engineering tasks like function-conditioned generation and variant screening.

주요 기능

01Efficient protein representation and embedding generation with ESM C

02Multimodal protein design across sequence, structure, and function using ESM3

03Automated 3D structure prediction and inverse folding capabilities

04Function-conditioned generation for targeted protein engineering

050 GitHub stars

06Flexible support for local model usage or cloud-based Forge API inference

사용 사례

01Optimizing protein variants through iterative chain-of-thought generation and refinement

02Extracting feature embeddings for protein classification, clustering, and similarity analysis

03Designing novel proteins with specific functional properties and structural constraints

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add workonthestreets/wolo esm

For use in Claude.ai and ChatGPT