Overview of the end-to-end pipeline used in the study. Images (20× objective magnification) of fluorescent immunohistochemistry-stained bone biopsy samples (DAPI, blue; CD34, purple; CD38, orange; CD117, green; CD61, yellow) were analyzed with multiple deep learning-based algorithms to explore the spatial dimension of human hematopoiesis. See Figures 1B and 2C in the article by Sarachakov et al that begins on page 2282.