New preprint on segmentation and quantification of podocyte morphology


Our work on automated analysis of podocytes is now online. For the first time we apply deep learning to super resolution microscopy images.

Together with the group of Prof. Benzing we came up with a method for segmentation and quantification of podocyte morphology. Podocytes - specialized cells in kidney - form intricate structures called foot processes. These protrusions in the cells form the filtration barrier of molecules from blood to urine. Their morphology is key for the impermeability of this barrier.

Our segmentation method performs on super resolution microscopy images of nephrin-stained tissue slides and marks separate foot processes. We then quantify their morphological parameters, such as area, perimeter, or circularity, and show how these parameters change in diseased kidney. Check out our preprint for more detail.