Pineoblastoma (PB) is an understudied, clinically aggressive embryonal tumor of the pineal gland diagnosed early in life. Due to its rarity, knowledge pertaining to the molecular and cellular basis of PB has remained modest. Likewise, preclinical models of PB are extremely scarce, stalling efforts to discover and credential curative therapies that reduce collateral damage to the developing central nervous system. To overcome the primitive understanding of PB tumor biology and identify clinically relevant molecular heterogeneity, we recently performed bulk genetic, epigenetic, and transcriptional profiling on a sizable cohort of primary patient tumor samples, defining four distinct subgroups: PB-RB1, PB-MYC/FOXR2, PB-miRNA1, and PB-miRNA2. Namesake PB-RB1 and PB-MYC/FOXR2 subgroups represent the primary diagnoses of PB during infancy and account for an unacceptably high mortality rate. Loss-of-function mutations targeting miRNA processing genes define the clinically heterogeneous PB-miRNA1 and PB-miRNA2 subgroups diagnosed in children and adolescents. An international consensus on the definition of these novel PB subgroups was included in the most recent update to the World Health Organization (WHO) Classification of Tumours of the Central Nervous System and has begun to change how PB is diagnosed clinically. In this research program, we aim to advance understanding of biologically and clinically distinct PB subgroups through deep investigation of their cellular, developmental, and pathological basis. We hypothesize that molecular subgroups of PB exhibit divergent cellular programs and originate from discrete progenitor states during pineal gland development. Disclosing cellular hierarchies and developmental origins of novel PB subgroups will (i) resolve oncogenic mechanisms; (ii) inform preclinical modeling strategies; (iii) implicate candidate dependencies; and (iv) enable evaluation of novel therapeutic strategies. This hypothesis will be tested in two conceptually and technically innovative Specific Aims that integrate a multidisciplinary and multispecies approach. In Aim #1, we will leverage a large single-nucleus transcriptome dataset generated from molecularly annotated PB patient tumors to discover cellular programs driving malignancy. In addition, we will utilize first-in-class atlases of the developing mouse and human pineal gland to define subgroup-specific cellular origins and reveal tumor-enriched molecular signatures, collectively informing pathogenic mechanisms and candidate therapeutic vulnerabilities. In Aim #2, we will deliver a series of novel, genetically engineered mouse (GEM) models representative of distinct PB subgroups, including comprehensive phenotypic and molecular characterization. Novel GEM models will validate suspected PB driver genes and serve as essential preclinical models for testing innovative treatment strategies targeting prominent cell identity programs. Successful execution of this research program will ove...