Speaker
Description
The polygenic architecture of human neurodiversity requires new maps to model the developmental continuum it underpins. While pluripotent stem cell and brain organoid modelling has begun to provide major insights into the pathogenesis of neuropsychiatric disorders, the focus is still on one or, at best, a few specific disorders at a time, mostly caused by highly penetrant genetic mutations and studied in isolation. The combination of single cell analysis of developmental trajectories through brain organoids modelling and advanced mathematical formalism has the potential to overcome the limitations of such classical “disease vs control” dichotomy and shift the paradigm towards the study of the continuous distributions of clinical and molecular phenotypes. For that, however, a significant scaling up is needed in the numbers of timepoints, cells and individuals profiled. Starting from anchor points from our single cell resolved reference of paradigmatic neurodevelopmental disorders (including Kabuki, Gabriele De Vries, Weaver, Williams Beuren, 7 dup and several others) at multiple time points and multiple individual per conditions, we developed a sophisticated mathematical approach to enable the continuous and probabilistic definition of trajectories, by extending nonlinear filtering and statistical learning methods. This unique combination of generative properties, computational and experimental alike, allows us to uncover hidden cellular microstates from noisy single-cell omics data.Through iterative refinement and experimental validation, this approach enables mapping, with unprecedented resolution, the molecular underpinnings of cellular dynamics onto clinical and neurobehavioral traits.
