AlphaGenome – now available through an API (pre-print, documentation). Start experimenting, and share your feedback and projects here!
Congratulations to the DeepMind team - this looks like an amazing resource for the scientific community. Are there any plans to enable access for users at for-profit companies (e.g. purchasing commercial license)?
Thanks so much for your question! We’re not able to discuss the product roadmap, but we’ll keep this community updated as we make updates and announcements.
Thanks for making API available. Now everyone can use it and projects from it.
I just wanted to say I really appreciate the documentation and tutorials, they made scoring variants and using the model quite straightforward.
One basic question I have is about the models that are available via the API. It looks like the only ones listed are trained on all folds or individual folds. Does this mean the distilled models used for variant effect prediction aren’t available through the API?
Hi @Sarthak_Ti,
The ALL_FOLDS model is our distilled model rather than one of our models trained on all folds. We’ll try and make this clearer in the documentation (thanks for pointing this out!)