BMI Seminar // Omer Bayraktar: Mapping the rules of glioblastoma with integrated single cell and spatial genomics

The malignant brain tumour glioblastoma multiforme (GBM) displays significant levels of cellular heterogeneity and plasticity that correlate with patient outcomes. However, we lack fundamental insight on how diverse malignant cell states arise and organise in GBM. Here, I will present GBM-space, a new collaborative effort to map the cellular and tissue architecture of GBM using multi-modal genomics. Deeply profiling 12 tumours across multiple sites, we created the most comprehensive single cell and spatial genomic map of inter- and intratumour heterogeneity to date. Surveying joint RNA and ATAC profiles of over 1 million cells, we expanded the description of recurrent malignant and TME cell states and resolved their gene regulatory networks. Using sequencing and imaging-based spatial transcriptomics, we discovered that malignant cell states regionally segregate into distinct spatial tissue niches in GBM and associate with distinct TMEs. We found that these niches recur across patients and tumour genotypes, reflecting a shared tissue organisation across GBM. Furthermore, we find that that the spatial organisation of malignant cells reflects their stereotyped cellular trajectories from neurodevelopmental-like to injury response and hypoxia states. Our efforts reveal novel spatial organisation of malignant cell states in glioblastoma across a dominant cellular trajectory.