esm

Generate per-residue and per-sequence embeddings with ESM-2 and predict structures via ESMFold.

18|2|Updated Feb 21, 2026
One-click install
npx skills add https://github.com/omar-A-hassan/medsci-agent --skill esm-omar-a-hassan
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: esm
Source: https://github.com/omar-A-hassan/medsci-agent/tree/main/.opencode/skills/esm
Command: npx skills add https://github.com/omar-A-hassan/medsci-agent --skill esm-omar-a-hassan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Researchers need scalable representations and structural insights for protein sequences to accelerate discovery.

Core Features & Use Cases

  • ESM-2 Embeddings: generate per-residue and per-sequence embeddings from amino acid sequences.
  • ESMFold: single-sequence structure prediction without MSAs.
  • Use Case: analyze a novel protein to obtain embeddings and a predicted structure for downstream analysis.

Quick Start

Provide a protein sequence to generate embeddings with ESM-2 and predict its structure with ESMFold.

Frequently Asked Questions about esm

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

FAQPage Schema
How do I generate protein embeddings from an amino acid sequence?

You can generate protein embeddings by providing an amino acid sequence to the ESM-2 model, which outputs per-residue and per-sequence representations suitable for downstream bioinformatics analysis.

What is single-sequence protein structure prediction without MSAs?

Single-sequence structure prediction is a process that uses ESMFold to predict a protein's 3D structure directly from an amino acid sequence, bypassing the need for multiple sequence alignments.

Do I need to install the fair-esm package to predict protein structures?

Yes, you need to install the fair-esm package and load pretrained models to run ESM-2 embedding workflows or ESMFold structure prediction on protein sequences.

Can I use ESM-2 to analyze a novel protein for downstream research?

Yes, you can analyze a novel protein by inputting its sequence to obtain ESM-2 embeddings and a predicted ESMFold structure, accelerating discovery for researchers and bioinformaticians.

What is the difference between per-residue and per-sequence embeddings?

Per-residue embeddings provide a representation for each individual amino acid in the sequence, while per-sequence embeddings provide a single aggregated representation for the entire protein.