Protein Tinder

The easy way to annotate proteins

A mobile-friendly, web-based platform for visualizing and annotating 3D protein structures using 3Dmol.js and NGL.js. Built with Django, this app supports protein annotation through intuitive gestures and interactive tools.

🚀 Use Cases

1. Swipe Annotations

To start with swipe annotations upload your compressed pdb folder (zip) and click on the folder name in the folder list.

👉 Swipe right — Mark as Correct
👈 Swipe left — Mark as Wrong
👇 Swipe down — Mark as Unsure

Annotations are saved to a MySQL database and can be exported as CSV from the web interface.

2. Domain Annotation

The upload process is the same for swipe/domain annotations. In the folder list click on `✏️Domains`.

🎯 Drag across the domain bar to define regions
🏷️ Name domains using the text field below
🖱️ Adjust domain boundaries by dragging

💾 Export zip:

CSV: One file per protein
FASTA: Each domain as a sequence
PDB: Each domain as a separate PDB

To download this zip: 1) Go to 👀Recap 2) make sure you are on the Domain Overview 3) Download using the download button in red: 📦 Download

Domain Number,Start Residue,End Residue,Predicted Domain
1,1,181,XD2
2,182,294,XD3

3. Domain Correction

To start with domain correction annotations upload your compressed pdb folder (zip) together with you domain annotations in csv format. The pdb filename and csv filename have to be identical except for the extension. After succesfull uploading, click on the folder name in the folder list to start domain correction annotations.

✅ Mark incorrect domain entries
📦 Upload .zip archives with .pdb and .csv files
☑️ Enable "Include architecture data?" during upload

📁 Each .csv must match its .pdb filename (e.g., protein_id.pdb and protein_id.csv).

🧪 How to Use

📝 Create an account
📤 Upload a .zip containing your .pdb files
✏️ Add an annotation title and description
⚙️ Annotate proteins using swipe or domain modes
📥 Export your annotations as .csv or .fasta

Note: The filename of each .pdb serves as its protein_id.

✨ Features

🧬 Interactive 3D protein viewer
📱 Mobile and desktop support
📊 Overview dashboard
🔐 User login & registration
👥 User-to-user interaction
💾 Export to CSV, FASTA, and PDB

🛠️ Tech Stack

Backend: Python, Django
Database: MySQL
Frontend: HTML, JavaScript, 3Dmol.js, NGL.js
Deployment: Docker

⚙️ Run Locally

git clone https://github.com/victornemeth/ProteinTinder.git
cd ProteinTinder
docker compose up -d --build

.env file should be created in root dir and look like this:

MYSQL_DATABASE=annotate_db
MYSQL_USER=annotate_user
MYSQL_PASSWORD=very_secure_password
MYSQL_ROOT_PASSWORD=another_very_secure_password

DJANGO_SECRET_KEY=another_another_very_secure_password
DJANGO_DEBUG=True  # For local use debug mode is recommended.

Visit localhost:8000 in your browser.

🌐 Note

The official website is deployed via NPM. This repository is for local development.