Highly recurrent, human syntenic CNAs characterize pediatric cancer models

Mouse models represent the gold standard for cancer-related preclinical in vivo trials. However, their molecular characterization has often been limited and lacks global DNA methylation profiling, which has emerged as a crucial classification tool for various human tumor entities. Therefore, the epigenetic tumor cell identity, proximity to respective human counterparts, and their underlying CNVs have mostly remained elusive. Here, we generate a comprehensive DNA methylation atlas of globally available mouse models for pediatric solid tumors, including an extensive, novel cohort of high-grade glioma models. We show that tumors cluster within dedicated tumor entities and exhibit previously suggested connections to putative cells of origin. Using tailor-made cross-species analysis, we highlight the epigenetic resemblance of analyzed models to the human disease and demonstrate entity-specific immune infiltration by custom deconvolution. Strikingly, we identified that specific mouse models harbor highly recurrent CNV signatures that are unique to the respective tumor subtype. We further show that these CNVs share syntenic regions with their respective human counterparts and occur reliably during multi-passage tumor evolution. Our results reveal epigenetic insights into a global conglomerate of mouse models for pediatric solid cancers and seminally showcase recurrent CNVs in murine tumors, granting long-awaited access to CNV research." Mouse sample were dissected and fresh frozen,cryoconserved or FFPE tissue was used for DNA extraction. The DNA was extracted and analyzed by the MM285 mouse methylation array (Illumina).