Google DeepMind’s AlphaGenome v0.10.0 adds plotting components for Atlas AVI scores and moves its model API, Colabs and protos into a stable v1 directory. That gives users new ways to inspect scores, while signalling that the previous locations are due to go away in a future release.
Google DeepMind Watch analysis
What happened
The release notes also add support for withnamesuffix and withmetadatacolumn for JunctionData, update documentation for the v1 directory structure and fix a bug with round-tripping scalar values. The team says the Atlas citation has been updated to a medRxiv pre-print. Read the AlphaGenome v0.10.0 release notes.
The move to v1 is the change to plan around: the notes say the current locations will be removed in a future release, but do not give a date. Users relying on existing paths should check their code and notebooks before that tidying-up becomes a migration task.
Why it matters
Atlas users get components for plotting feature contributions and scores, while JunctionData gains two options for naming and metadata columns. The scalar fix is less showy, but reliable round-tripping is the sort of quiet improvement that saves a data workflow from developing a personality of its own.
Our read
This is a practical update for people working with AlphaGenome data, not a new claim about what the model can predict. The v1 move deserves attention now: check which paths your work depends on, and treat the old locations as temporary even though the removal date is not yet specified.
What to watch
- When Google DeepMind will remove the current API, Colab and proto locations.
- How the new plotting components fit into Atlas users’ analysis workflows.
- Whether the JunctionData additions and scalar fix address issues in your own pipelines.
Discussion spark: Should projects set a firm removal date when they move APIs to a new stable directory, or is an advance warning without a date enough?
Sources and evidence
- v0.10.0 (8 October 2026, 13:56 UTC)
not affiliated with, endorsed by, or operated by Google or Google DeepMind