Lectures and seminars Spatiotemporal approaches to dissecting brain cell state emergence, dynamics, and modulation
Welcome to a seminar by Dr. Gioele La Manno, Assistant Professor, Swiss Federal Institute of Technology Lausanne.
Brain development involves complex cellular heterogeneity, with states that change across time and space. Understanding these processes requires approaches that can capture how cellular states transition, organize spatially, and respond to signals. Such approaches are the more useful the more they are able dissect such a complex machinery and the more they offer opportunities to reveal how disruptions contribute to disease. I will present three research lines that are motivated by this need to advance our fine understanding of brain developmental processes.
Host: Ana Teixeira, Department of Physiology and Pharmacology, KI.
First, focusing on dynamics, I will describe how we elevated RNA velocity estimation into a statistical inferential framework for answering well-posed biological questions. I will introduce VeloCycle, a heuristics-free Bayesian method that exploits the consistency between vector fields and manifold geometry while correctly quantifying uncertainty. We validate that VeloCycle detects quantitative differences and captures speed variations between prosomers and in Perturb-seq screens. I will also highlight our latest generalization, which decomposes velocity into interpretable biological components along which statistical inference can be conducted.
Second, I will present our systematic mapping of neural progenitor organization in the embryo using large-scale HybISS, enhanced by two deep learning tools: Spotiflow and PointillHist. This work charted hundreds of distinct cell states, with PointillHist learning histological patterns to build a comprehensive atlas of embryonic brain organization. This atlas provides a reference framework for interpreting perturbation outcomes, ranging from subtle adaptations to catastrophic malformations. Specifically, I will show that Shh organizer ablation triggers region-specific cellular responses, while folate deficiency exposes unexpected vulnerabilities in high Folr1-expressing radial glia of cortical hem organizers.
Finally, moving beyond transcriptomics, I will focus on key cellular components that modulate how cells respond to signals: membrane lipids. I will present uMAIA, a platform enabling large-scale lipidomic profiling and construction of the first comprehensive Lipid Brain Atlas. This resource reveals “lipizones”—spatially organized metabolic territories that define functional anatomy beyond cellular composition. Lipizones capture connectivity patterns, uncover oligodendrocyte heterogeneity axes in white matter, and reveal biochemical zonation in the choroid plexus and ventricular walls. We further show that this lipidomic architecture adapts dynamically to physiological demands during pregnancy, opening a new axis for exploring signaling modulation in brain development.
Gioele La Manno leads the Laboratory of Brain Development and Biological Data Science at EPFL, where he produces unique datasets and develops methods to understand neural development. He earned his PhD with Sten Linnarsson at Karolinska Institute, where he categorized brain progenitor cells subtypes in mammals and invented RNA Velocity Analysis. In 2018, he started his lab at EPFL directly after his Ph.D., and since 2023, he has been a Tenure Track Assistant Professor. His recent work spans spatial transcriptomics, data science, and lipidomics, where he co-discovered lipid-based states. He has been awarded Starting grants from both the ERC and the Swiss National Foundation and was recognized him multiple prizes, including the Vasco Sanz, EMPIRIS, SIB, and Chorafas awards.