Add example data for antibody-antigen docking models use case

[gcroci2]

• Ask A. Bonvin for data (or at least samples of them)
  • They are in the process of publishing another study/dataset. They are also working on these systems and plan to generate a larger dataset.
  • Evaluate if the complete dataset, which we don't need yet in this phase, can be used for a future phase of the project, i.e., running a satisfying training for a relevant research question
• Upload an example dataset somewhere accessible (e.g., Zenodo)
• For ranking the models
• https://data.sbgrid.org/dataset/686/
  • Chat with Cunliang
    • They seem to be 2 chains-things. What's the difference with PPIs we have and why we can't support already such complexes?
    • Where are the native structures?
    • Are 16 complexes enough to train a DL model?
    • Brainstorm about the use case
  • What about infrastructure for this?

[github-actions]

This issue is stale because it has been open for 30 days with no activity.