---
title: UniProt quickstart
description: >-
  Search for a protein in UniProt and chain the sequence into structure
  prediction.
tags:
  - uniprot
  - quickstart
---

You want a protein's sequence and annotations to start a structure pipeline; this searches [UniProt](/docs/uniprot), then chains the sequence straight into ESMFold structure prediction.

## UniProt Quickstart

## Step 1: Search

In chat, ask for a protein:

```
Search UniProt for human EGFR kinase domain
```

The platform queries the UniProt API and returns matching protein entries.

## Step 2: Review Annotations

Results include:
- **Protein name** and gene symbol
- **Accession ID** (e.g., P00533)
- **Organism** and taxonomy
- **Function** description
- **GO terms** — molecular function, biological process, cellular component
- **Sequence** — full amino acid sequence
- **Cross-references** — PDB entries, Pfam domains, InterPro families

## Step 3: Chain into Structure Prediction

Ask the platform to predict the structure:

```
Fold this sequence with ESMFold
```

Or chain directly in one prompt:

```
Fetch EGFR kinase domain from UniProt and fold it with ESMFold
```

## Step 4: Continue the Pipeline

After folding, you can:
- **Predict binding sites**: Run P2Rank on the predicted structure
- **Run a similarity search**: Send the sequence to BLAST
- **Build a target passport**: Generate a comprehensive target profile

## Next Steps

- Ask for the whole chain in one message (*"Fold P00533 and find its binding pockets"*) and the [orchestrator](/docs/how-orchestration-works) runs UniProt → ESMFold → P2Rank for you
- Search for homologs with [BLAST](/docs/blast)
- Generate embeddings with [ESM-3](/docs/esm3)

## Read next