UCSC adds DeepMind AlphaGenome Variant Impact scores for all 8.8B single-base changes
anshulkundaje · x · 2026-09-11
UCSC Genome Browser has added a new hg38 track of Google DeepMind's AlphaGenome Variant Impact (AVI) scores, with precomputed scores for roughly 8.8 billion possible single-base substitutions across the human genome.
Key points:
- AVI combines AlphaGenome's predictions of regulatory effects (expression, splicing, chromatin accessibility, TF binding across hundreds of cell types) with AlphaMissense predictions for protein-altering changes, outputting a single PHRED-scaled score (10 = top 10% genome-wide, 30 = top 0.1%).
- Unlike most prediction scores, AVI covers non-coding variants too — e.g., it flags TERT promoter mutations common in cancer.
- Each alternate allele gets its own subtrack for direct locus-level browsing.
More from Research
- Researcher warns CHI is about to be flooded by AI slop papers — IanArawjo · 2026-09-11
- Rumor: a counterexample already exists, Lean formalization being finalized — lpachter · 2026-09-11
- CiNet's 11th international conference on biological to general AI opens registration — kaixhin · 2026-09-11
- Looped Transformers Hit a Natural Ceiling: Most Test-Time Scaling Is Environment Interaction — generativist · 2026-09-11
- Classic free book 'From Python to NumPy' now readable online with an AI tutor — burkov · 2026-09-11
- Author's caveat: stroke model's Dice 0.46 is an upper bound from outcome-fitted parameters — maier_ak · 2026-09-11