Researchers used AI to identify a DNA activation initiator sequence, reporting a model that detects the initiator in roughly 60% of human genes after analyzing about 500,000 DNA sequences. The work decodes a key genetic “switch” that helps control whether genes turn on. The advance can improve mechanistic interpretation of harmful mutations by connecting sequence context to gene activation behavior. While the study is primarily computational and model-driven, it provides a concrete target for follow-up functional validation. The output is relevant for translational genomics, where linking variants to regulatory consequences is often the bottleneck in variant interpretation workflows.