AlphaFold2 pipeline in Cloud via Google Colab
The Colab is an online Colaboratory Google Research platform, free to use by anyone within the usage limits. The initiative provides users with computing resources (both CPU and GPU) via the Jupyter notebook interface in the browser. The latter supports free software, open standards, facilitate collaboration, and provides web services for interactive computing across all programming languages. For the Colab framework, Python is the preferred coding language. The notebooks broadly meet the need for an interactive and user-friendly working environment. In a single document, you can combine blocks of executable code, human-readable comments in rich text along with the results of analysis, including graphs, images, HTML, LaTeX and more.
ColabFold notebook: AlphaFold2 using MMseqs2
The AlphaFold2.ipynb notebook is available online at https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb. The current version allows modeling of both monomeric and multimeric protein structures using AlphaFold2 and Alphafold2-multimer, respectively. When you shorten the URL in your browser’s search box to ~ColabFold/, you’ll find a list of other more specialized notebooks, such as for complexes.
Using the Alphafold notebook to predict the structure of a protein is straightforward.
- First, copy-paste the amino acid sequence in one-letter notation to the
The example above contains the amino acid sequence in single letter notation for a small globular protein G. For starters, you can use this sequence.
- Then find “Runtime” in the top menu, click to expand available options, and press “Run all”.
The detailed instructions are provided directly in the bottom part of the notebook, including Limitations, Issues, and Troubleshooting.
Overall structure of the ColabFold notebook (AlphaFold2)
• Input protein sequence(s) query_sequence: jobname: use_amber: use_templates: save_to_google_drive: • Advanced settings msa_mode: model_type: pair_mode: num_recycles: • Install dependencies • Run Prediction • Display 3D structure rank_num: color: show_sidechains: show_mainchains: • Plots • Package and download results
ColabFold by Boston Protein Desin @YouTube
To find the scientific insights of AlphaFold modeling as well as to learn useful technical tips & tricks of using the ColabFold implementation, please watch the “ColabFold - Making protein folding accessible to all via Google Colab!” movie recorded on August 4th, 2021 presented by Sergey Ovchinnikov and Martin Steinegger, hosted by Chris Bahl.