Massive chromosome-wide gains and losses—not single-gene mutations—kick-start Group 3/4 medulloblastoma while the famous MYC and MYCN amplifications sneak in later as minor but dangerous subclones. A new Nature study that pairs snATAC-seq with spatial transcriptomics maps these prenatal “copy-number quakes” and shows how single-cell epigenomics can spot the tumour’s earliest fault lines. Dive into the blog to see why rethinking cancer’s first hit could reshape risk prediction and therapy.