Chromosome-Scale Shifts, Not Single Genes, Ignite Medulloblastoma
Single-cell and spatial multi-omics reveal prenatal copy-number “earthquakes” that set the stage for later MYC/MYCN flare-ups
Why researchers should be excited
A new Nature study by Okonechnikov et al. shifts the narrative in childhood brain tumours: sweeping copy-number variation (CNV), not headline oncogenes, appears to spark Group 3/4 medulloblastoma. By combining single-nucleus RNA-seq, single-nucleus ATAC-seq (snATAC-seq) and single-cell spatial transcriptomics across 20 tumours, the authors show that broad chromosome gains and losses arise in utero, while familiar drivers such as MYC, MYCN or enhancer-hijacked PRDM6 appear later, in small but aggressive subclones.
Clonal proliferation and differentiation gradients are independent of oncogene expression
Key revelations from the atlas
CNVs fire first. Every tumour’s founder clone carried chromosome-wide events—loss of chr 10, gain of chr 17q or 7—visible in both snRNA- and snATAC-derived CNV profiles.
Oncogenes arrive late—and tiny. Amplifications of MYC, MYCN or enhancer-driven PRDM6 surfaced only in descendant branches, often representing a minor-clone fraction in the primary mass.
Tumours start before birth. Molecular-clock modelling dates the first CNVs to the first gestational trimester in roughly one-quarter of cases, with the remainder emerging during infancy—implying a long silent latency.
Spatial mixing dominates. Single-cell spatial maps revealed largely intermingled subclones; occasional MYC- versus MYCN-enriched “islands” foreshadowed which clone would dominate at relapse.
Methodological sparks—and why snATAC-seq changed the game
snATAC nailed the CNV calls. When CNV profiles were cross-checked against bulk methylation data, those inferred from snATAC always matched the correct sample, whereas snRNA produced several false-positive matches.
Rare oncogene clones rescued. Many tumours appear “driver-less” in bulk sequencing because MYC/MYCNclones are tiny. snATAC exposed these rare subclones, revealing that each tumour actually harboured an oncogene-positive branch—crucial information for relapse risk.
Deep sequencing, deeper insight. Tripling snATAC coverage in three MYC cases recovered up to 40 % of somatic SNVs and assigned many to specific subclones—granularity unattainable with the transcriptome alone.
Blueprint for future spatial epigenomics. Integrating this “zoom-lens” epigenetic view with high-resolution spatial chromatin assays (e.g., spatial ATAC-seq) will allow researchers to track both CNV architecture and regulatory rewiring in situ.
Why it matters beyond medulloblastoma?
These findings overturn a long-held assumption: in paediatric tumours the decisive first hit can be a chromosome-scale quake, not a single-gene mutation. Therapies that chase focal drivers risk sparing the founding CNV clone, leaving a reservoir for relapse. Early, spatially resolved CNV mapping therefore offers a new path for risk stratification and pre-emptive intervention.
Open questions for the field
Can prenatal CNV signatures be detected non-invasively (for example, in circulating fetal DNA)?
What micro-environmental cues let a late MYC subclone overtake its CNV-defined siblings at relapse?
Do similar CNV-first trajectories drive other paediatric solid tumours?
From insight to action—why spatial epigenomics matters
Okonechnikov et al. show that combining single-cell and spatial views can rewrite a tumour’s timeline. The next leap is to overlay copy-number states and regulatory chromatin simultaneously at cellular resolution—an emerging capability of spatial epigenomic approaches such as spatial ATAC-seq. By mapping both the structural and regulatory genome directly on tissue, researchers can pinpoint which early clones are poised for domination and which micro-niches nurture their rise, paving the way for truly stage-matched interventions.
Further reading: “Oncogene aberrations drive medulloblastoma progression, not initiation,” Nature (online 2025).
Oncogene aberrations drive medulloblastoma progression, not initiation | Nature
