DeepMind published AlphaGenome, a long‑sequence DNA model that predicts regulatory effects of non‑coding variants across up to one million base pairs. The Nature paper shows the model can infer multiple molecular properties—gene starts/ends, splicing, accessibility—at scale and in seconds, addressing a key bottleneck in interpreting regulatory DNA. AlphaGenome is pitched as a tool to prioritize non‑coding variants from GWAS and rare‑disease studies and to accelerate therapeutic target discovery and synthetic regulatory design. The model’s ability to connect distant regulatory elements could reshape variant interpretation pipelines and target identification for genomics‑driven drug discovery.