Notebooks
W
Weights and Biases AlphaFold With W&B Align, Fold, Log

AlphaFold With W&B Align, Fold, Log

tableswandb-examplescolabs
alph-notebooks/wandb-examples / AlphaFold_with_W&B_Align,_Fold,_Log.ipynb

Open In Colab

Weights & Biases

This Colab notebook allows you to easily predict the structure of a protein using a slightly simplified version of AlphaFold v2.0.

This notebook was modified from the original produced by DeepMind to add additional W&B logging features.

You can read more in this Report and view an example protein below.

The text below is from the original.

Differences to AlphaFold v2.0

In comparison to AlphaFold v2.0, this Colab notebook uses no templates (homologous structures) and a selected portion of the BFD database. We have validated these changes on several thousand recent PDB structures. While accuracy will be near-identical to the full AlphaFold system on many targets, a small fraction have a large drop in accuracy due to the smaller MSA and lack of templates. For best reliability, we recommend instead using the full open source AlphaFold, or the AlphaFold Protein Structure Database.

Please note that this Colab notebook is provided as an early-access prototype and is not a finished product. It is provided for theoretical modelling only and caution should be exercised in its use.

Citing this work

Any publication that discloses findings arising from using this notebook should cite the AlphaFold paper.

Licenses

This Colab uses the AlphaFold model parameters and its outputs are thus for non-commercial use only, under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license. The Colab itself is provided under the Apache 2.0 license. See the full license statement below.

More information

You can find more information about how AlphaFold works in our two Nature papers:

FAQ on how to interpret AlphaFold predictions are here.

0. Install & import packages

[ ]
[ ]
[ ]
[ ]
[ ]

#1. Enter a sequence and make a prediction

Please paste the metadata and sequence of your protein in the code cell box below, then run it and the remaining cells via Runtime > Run after. You can also run the cells individually by pressing the Play button on the left.

Note that the search against databases and the actual prediction can take some time, from tens of minutes to hours, depending on the length of the protein and what type of GPU you are allocated by Colab (see FAQ below).

Sample sequences

We've provided a few example sequences in case you want to get started quickly. You can also search for proteins in one of these databases:

[ ]
[ ]
[ ]
[ ]
[ ]

Interpreting the prediction

Please see the AlphaFold methods paper and the AlphaFold predictions of the human proteome paper, as well as our FAQ on how to interpret AlphaFold predictions.

2. Log to W&B: Upload the sequence, 3D structure, and results

[ ]

FAQ & Troubleshooting

  • How do I get a predicted protein structure for my protein?
    • Click on the Connect button on the top right to get started.
    • Paste the amino acid sequence of your protein (without any headers) into the “Enter the amino acid sequence to fold”.
    • Run all cells in the Colab, either by running them individually (with the play button on the left side) or via Runtime > Run all.
    • The predicted protein structure will be downloaded once all cells have been executed. Note: This can take minutes to hours - see below.
  • How long will this take?
    • Downloading the AlphaFold source code can take up to a few minutes.
    • Downloading and installing the third-party software can take up to a few minutes.
    • The search against genetic databases can take minutes to hours.
    • Running AlphaFold and generating the prediction can take minutes to hours, depending on the length of your protein and on which GPU-type Colab has assigned you.
  • My Colab no longer seems to be doing anything, what should I do?
    • Some steps may take minutes to hours to complete.
    • If nothing happens or if you receive an error message, try restarting your Colab runtime via Runtime > Restart runtime.
    • If this doesn’t help, try resetting your Colab runtime via Runtime > Factory reset runtime.
  • How does this compare to the open-source version of AlphaFold?
    • This Colab version of AlphaFold searches a selected portion of the BFD dataset and currently doesn’t use templates, so its accuracy is reduced in comparison to the full version of AlphaFold that is described in the AlphaFold paper and Github repo (the full version is available via the inference script).
  • What is a Colab?
  • I received a warning “Notebook requires high RAM”, what do I do?
    • The resources allocated to your Colab vary. See the Colab FAQ for more details.
    • You can execute the Colab nonetheless.
  • I received an error “Colab CPU runtime not supported” or “No GPU/TPU found”, what do I do?
    • Colab CPU runtime is not supported. Try changing your runtime via Runtime > Change runtime type > Hardware accelerator > GPU.
    • The type of GPU allocated to your Colab varies. See the Colab FAQ for more details.
    • If you receive “Cannot connect to GPU backend”, you can try again later to see if Colab allocates you a GPU.
    • Colab Pro offers priority access to GPUs.
  • Does this tool install anything on my computer?
    • No, everything happens in the cloud on Google Colab.
    • At the end of the Colab execution a zip-archive with the obtained prediction will be automatically downloaded to your computer.
  • How should I share feedback and bug reports?
    • Please share any feedback and bug reports as an issue on Github.

Related work

Take a look at these Colab notebooks provided by the community (please note that these notebooks may vary from our validated AlphaFold system and we cannot guarantee their accuracy):

License and Disclaimer

This is not an officially-supported Google product.

This Colab notebook and other information provided is for theoretical modelling only, caution should be exercised in its use. It is provided ‘as-is’ without any warranty of any kind, whether expressed or implied. Information is not intended to be a substitute for professional medical advice, diagnosis, or treatment, and does not constitute medical or other professional advice.

Copyright 2021 DeepMind Technologies Limited.

AlphaFold Code License

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at https://www.apache.org/licenses/LICENSE-2.0.

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Model Parameters License

The AlphaFold parameters are made available for non-commercial use only, under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license. You can find details at: https://creativecommons.org/licenses/by-nc/4.0/legalcode

Third-party software

Use of the third-party software, libraries or code referred to in the Acknowledgements section in the AlphaFold README may be governed by separate terms and conditions or license provisions. Your use of the third-party software, libraries or code is subject to any such terms and you should check that you can comply with any applicable restrictions or terms and conditions before use.

Mirrored Databases

The following databases have been mirrored by DeepMind, and are available with reference to the following: