Genetic sequencing, an essential method in life science, is mainly driven by fluorescence imaging. In the advent of spatial imaging and in situ sequencing directly on cells and human tissues, we aim to apply latest deep learning and microscopic automation for translational medicine and molecular pathology. This includes fast optical imaging, single cell sequencing, machine learning and describes well the foci of the intelligent imaging group to study and improve patient care at Charité – BIH Digital Health Center.
Automated light sheet microscopy
Advanced automated light sheet microscopy is the mode of choice for extremely fast fluorescence 3D imaging of a cellular context. Dual top geometries and oblique light sheet microscopy are developed for the acquisition of spatial genetics in micro tissues and organoids. Especially the massive imaging over space or time requires for intelligent classification or deep learning algorithms, to selectively save and identify relevant genetic information in volumetric heterogenous patient biopsies.
Human single cell sequencing
Our group engineers single cell and nuclei RNA/ATAC omics sequencing libraries from patient tissues or derived organoids. Currently, we concentrate on single nuclei protocols for lung and pancreas samples preventing high autolytic capacity after tissue resection. Novel cell types and transcriptomics data are embedded in the Human Cell Atlas initiative. In general, we are employing single cell sequencing to understand different disease entities on a mechanistic cellular level, including COVID-19 since 2020.
Deep tissue learning
Artificial deep neural networks can learn to recognize and reconstruct patterns in single cell data. This allows a de novo identification of functional gene sets, master and housekeeping genes. Ultimately, tissue images should and will contain all features learnt in deep neural networks to enable seamless disease prediction and subsequent therapy management.