Event Details:
The first Monday of each month, the Knight Initiative for Brain Resilience will host monthly seminars to bring together awardees, affiliated professors and students for a series of 'lab meeting' styled talks. Two speakers will discuss their brain resilience research, experiences in the field, and answer questions about their work.
Sanket Gupte, Stanford University
AI methods for automated quantification of organelle phenotyes in CryoET data
Cryo-electron tomography (CryoET) is a powerful technique for imaging subcellular structures in 3D. Recent advances in this field have enabled the study of increasingly complex biological systems through the characterization of ultrastructural features at the nanometer scale. Quantitative analyses of CryoET data can be used to identify biomarkers of diseases, enabling the discovery of novel mechanistic insights and informing the development of potential treatments. However, manual annotation of dense volumes is impractical at the scales required to extract these insights. To overcome this limitation, we've developed CryoVIT, a deep learning method for accurately segmenting pleomorphic structures such as mitochondria while requiring very few ground truth labels. Our approach enables large-scale quantitative and morphological studies of cellular features in CryoET data. We demonstrate this ability by analyzing mitochondrial structures in neurons differentiated from iPSC of Huntington’s Disease patients.
Gong-Her Wu, Stanford University
Revealing Pathological Cellular Indicators of Huntington's Disease Using iPSC-Derived Neurons, Brain Cells, and Tissues via Cryogenic Electron Tomography
Unveiling early-stage Huntington's Disease (HD) pathology is critical for devising effective treatments. Our study used a variety of electron imaging tools to identify potential cellular markers in HD cells. We tackled this by cultivating iPSC-derived neurons on EM grids, vitrifying them by plunge or high pressure freezing, using focused ion beams scanning electron microscopy to generate a volume density by slice and view, preparing a thin slice of lammela and then utilizing cryogenic electron tomography (cryoET) to generate tomograms followed by 3D reconstruction. We identified novel biomarkers: mitochondria with enlarged granules and double membrane-bound electron-dense sheet-like aggregates within diseased neuronal cells. Strikingly similar traits emerged in HD (BACHD) primary neurons and R6/1 mouse brain tissue. To quantify these traits, we employed artificial intelligence and expanded our research to quantify abnormal granules in the mitochondria of these samples. The observed cellular structure features can be used as pathological markers when we apply therapeutic intervention reagents.