Machine Intelligence in the Life Sciences

 

Our Machine Intelligence in the Life Sciences group develops AI methods for analyzing high-resolution microscopy data. Our focus is on how molecular processes – particularly the structure, dynamics, and interactions of proteins and protein complexes – can be better visualized and quantitatively assessed using modern nanoscopy and cryo-electron microscopy techniques. To this end, we employ machine learning methods to analyze complex signals from techniques such as STED, MINSTED, and MINFLUX microscopy as well as cryo-electron tomography more reliably, to localize individual molecules more precisely, and to link the obtained information with learnedprotein structure models such as AlphaFold.

The group at MPI-NAT thus expands upon our previous and ongoing work at the University of Göttingen. There, our focus is primarily on AI-supported analysis of cells, tissues, organs, and entire organisms using fluorescence and electron microscopy. Our research at the MPI complements this approach at the molecular level: it aims to contribute to the direct investigation of biological processes on the nanometer scale within their cellular context. In this way, our approach combines state-of-the-art microscopy, machine learning, and structural biology while simultaneously strengthening existing collaborations between the University of Göttingen, the MPI-NAT, and partners from the biotechnology industry.