afusion 0.2.1

A Streamlit app for running AlphaFold 3 predictions

🔬 AFusion: AlphaFold 3 GUI

AFusion is a user-friendly graphical interface designed to simplify the process of generating input JSON files and running AlphaFold 3 predictions. It streamlines the setup and execution of AlphaFold 3, making it accessible to users who prefer a GUI over command-line interactions.

Demo site (generate input JSON files ONLY)

Table of Contents

• Features
• Prerequisites
• Installation
• Usage
  • Launching AFusion
  • Using the GUI
• Screenshots
• License
• Acknowledgements

Features

• Intuitive Interface: Easily configure job settings, sequences, and execution parameters through a clean and modern GUI.
• Entity Management: Add multiple entities (Protein, RNA, DNA, Ligand) with support for modifications, MSA options, and templates.
• Dynamic JSON Generation: Automatically generates the required JSON input file for AlphaFold 3 based on user inputs.
• Integrated Execution: Run AlphaFold 3 directly from the GUI with customizable Docker execution settings.
• Visual Feedback: Provides command output within the interface for monitoring and debugging.

Prerequisites

Before using AFusion, ensure that you have the following:

1. Docker Installed: Docker is required to run AlphaFold 3. Install Docker from the official website.

2. AlphaFold 3 Installed: AFusion requires AlphaFold 3 to be installed and set up on your system. Follow the installation instructions provided in the AlphaFold 3 GitHub Repository to deploy AlphaFold 3.

3. Python 3.10 or Higher: AFusion is built with Python and requires Python 3.10 or higher.

4. Streamlit Library: Install Streamlit to run the GUI application.

   pip install streamlit
   

Installation

1. Clone the Repository: Clone or download the AFusion source code from this repository.

   git clone https://github.com/yourusername/afusion.git
   

2. Navigate to the Project Directory

   cd afusion
   

Usage

Launching AFusion

1. Start the Streamlit App

   From the project directory, run:

   streamlit run app.py --server.fileWatcherType=none
   

2. Access the Application

   • The application will launch, and Streamlit will provide a local URL (e.g., http://localhost:8501).
   • Open the provided URL in your web browser to access AFusion.

Using the GUI

Find more about input in here.

1. Welcome Page

• Logo and Introduction: You'll see the AFusion logo and a brief description.
• Navigation Sidebar: Use the sidebar on the left to navigate to different sections of the app.

2. Job Settings

• Job Name: Enter a descriptive name for your job.
• Model Seeds: Provide integer seeds separated by commas (e.g., 1,2,3).

3. Sequences

• Number of Entities: Select how many entities you want to add (Proteins, RNA, DNA, Ligand).
• Entity Details: For each entity:
  • Entity Type: Select the type (Protein, RNA, DNA, Ligand).
  • Entity ID: Provide an identifier for the entity.
  • Sequence Input: Enter the sequence information.
  • Modifications: Optionally add modifications with their types and positions.
  • MSA Options: Choose MSA generation options and provide MSA data if applicable.
  • Templates: Optionally add template data with mmCIF content and indices.

4. Bonded Atom Pairs (Optional)

• Add Bonds: Check the box to add bonded atom pairs.
• Define Bonds: For each bond, provide details for the first and second atoms, including entity IDs, residue IDs, and atom names.

5. User Provided CCD (Optional)

• User CCD Input: Paste or enter custom CCD data in mmCIF format.

6. Generated JSON

• Review JSON Content: The application generates the JSON input file based on your entries. You can review it here.

7. AlphaFold 3 Execution Settings

• Paths Configuration:

  • AF Input Path: Specify the path to the AlphaFold input directory (e.g., /home/user/af_input).
  • AF Output Path: Specify the path to the output directory (e.g., /home/user/af_output).
  • Model Parameters Directory: Provide the path to the model parameters directory.
  • Databases Directory: Provide the path to the databases directory.

• Execution Options:

  • Run Data Pipeline: Choose whether to run the data pipeline (CPU-intensive).
  • Run Inference: Choose whether to run inference (requires GPU).

8. Run AlphaFold 3

• Save JSON File: Click the "Save JSON File" button to save the generated JSON to the specified input path.
• Run AlphaFold 3 Now: Click the "Run AlphaFold 3 Now ▶️" button to execute the AlphaFold 3 prediction using the Docker command.
  • Docker Command: The exact Docker command used is displayed for your reference.
  • Command Output: Execution output is displayed within the app for monitoring.

Screenshots

License

This project is licensed under the GPL3 License - see the LICENSE file for details.

Acknowledgements

• AlphaFold 3: This GUI is designed to work with AlphaFold 3 by DeepMind.
• Streamlit: AFusion is built using Streamlit, an open-source app framework for machine learning and data science teams.
• Contributors: Waiting for more!

---

If you encounter any issues or have suggestions for improvements, please open an issue or submit a pull request.

Happy Folding! 🧬