Colab: ColabFold on Google Colab

Another possibility to access the AlphaFold models is via ColabFold. This is a GitHub repository maintained by Sergey Ovchinnikov (Harvard University), containing multiple Colab notebooks with AlphaFold2 implementations, relaxation, ESMFold, RoseTTAFold2 and OmegaFold.

Predict the same complex as before (for example, SARS-CoV-2 RBD with VHH-E) via the ColabFold notebook (this requires a Google account). In the AlphaFold2_mmseqs2 notebook, you can add a query sentence, where you separate different chains with a “:” sign. Then, select Runtime > Run All. You can then monitor the prediction progress, and a .zip file with results will be automatically downloaded. You can later compare these results to the results that you obtained from the HPC.