AlphaFold Tutorial using ColabFold

Welcome to the AlphaFold tutorial where we will use a modified version of ColabFold to predict protein folding structures using AlphaFold2 and MMseqs2. This tutorial is designed to help you gain a deeper understanding of the protein folding prediction process, and to enable you to predict protein structures with ease using ColabFold.

Introduction

ColabFold is a protein folding prediction tool based on Google DeepMind's AlphaFold and utilizes MMseqs2 for sequence alignments and templates. With ColabFold, you can easily predict the protein folding structure based on the amino acid sequence.

Running the Tutorial

The tutorial is a Jupyter notebook that can be run on UCloud. To run each cell, press Shift + Enter while inside the cell or press the triangular play button above. There will be questions included throughout the workshop to help you gain a deeper understanding of what you are calculating.

Learning Objectives

After completing this workshop, you will be able to:

• Describe how to use ColabFold to predict protein structures.
• Explain and interpret the results generated using ColabFold.
• Use ColabFold to predict the protein structure of any specific protein of interest.

Pre-Analysis Technical Requirements

Before beginning the analysis, it is important to ensure that the following technical requirements are met:

1. The tutorial can be executed within a Docker container (as detailed in the Dockerfile), or through the UCloud platform.
2. The Jupyter Notebook has only been tested on JupyterLab version 3.1.4, therefore it is recommended to use this version for optimal performance.
3. To achieve optimal and efficient prediction results, it is strongly recommended to allocate at least one GPU for the analysis.
4. To initiate the analysis, simply open the AlphaFold2.ipynb file and run the cells as instructed.
5. A clear understanding of protein folding and AlphaFold2's working principles is required before starting the analysis.
6. Make sure you have read the manuscript and familiarized yourself with the ColabFold GitHub page for a comprehensive understanding of the tool.
7. It is advisable to have basic understanding of Jupyter Notebook and its functionality for seamless execution of the analysis.

Getting Started

You can access ColabFold through UCloud. This integration allows you to use the tool, but before beginning the ColabFold workshop, you need to download the workshop as a Jupyter Notebook from GitHub and upload it to your designated folder on UCloud. To initiate this process, please follow the steps below:

1. Log in to UCloud.
2. Choose a workspace that has GPU resources for optimal performance.
3. Search for ColabFold in the Apps section.
4. Choose the appropriate machine type and select a GPU, such as u2-gpu-1. The use of a single GPU is generally sufficient, though larger sequences may require additional computational power.
5. Select a personal folder to use, ensuring that you have uploaded the Jupyter Notebook from GitHub beforehand.
6. Click on "Submit" and then "Open interface" on the following page.
7. Prior to proceeding, ensure you've downloaded the Jupyter Notebook labeled AlphaFold2.ipynb from this repository. Then, upload the notebook to your ColabFold session using the left-hand menu.
8. Proceed with the ColabFold workshop, which will guide you through the process of predicting protein structures based on amino acid sequences.

Download AlphaFold2.ipynb using Terminal

You can use the git command to easily download the Jupyter Notebook from this GitHub repository using the terminal. Here's a step-by-step guide on how to do this:

1. Clone the Repository: Open your terminal and navigate to the directory where you want to store the downloaded repository. Use the following command to clone the repository:

git clone https://github.com/hds-sandbox/AlphaFold_Workshop.git

You can find the URL by clicking on the green "Code" button on the repository's GitHub page.

2. Copy the Notebook You can simply copy the notebook file to your desired location using the cp command. For example:

cp AlphaFold2.ipynb /path/to/destination

Replace /path/to/destination with the path where you want to save the notebook on your local machine.

That's it! You've successfully downloaded the Jupyter Notebook from the GitHub repository using the terminal. Make sure you have Git installed on your system, and if not, you can download it from the official Git website: https://git-scm.com/downloads.

References

Mirdita M, Schütze K, Moriwaki Y, Heo L, Ovchinnikov S, Steinegger M. ColabFold: Making protein folding accessible to all. Nature Methods, 2022.