Fig. 1: Assay workflow. | npj Artificial Intelligence

Fig. 1: Assay workflow.

From: AI driven 3D subcellular RPE map discovers cell state transitions in establishment of apical-basal polarity

Fig. 1: Assay workflow.The alternative text for this image may have been generated using AI.

A Immature iRPE cells from 16 lines, each expressing a different organelle protein marker tagged with GFP, were grown for four weeks in the presence of PGE2, a primary cilium enhancer, or HPI4, an inhibitor of primary cilium trafficking, to promote or inhibit iRPE cell apical/basal polarity. B Immature iRPE cells were seeded in glass-bottom 96-well plates. Seven days after cell seeding, PGE2 or HPI4 treatment was started to respectively promote or prevent cell polarization. Once a week, a plate was fixed in 4% PFA to preserve RPE cell states at different timepoints of RPE polarization. The first timepoint (week 1) represents a baseline since treatment with PGE2 and HPI4 started after the first collection. Cells were stained with phalloidin-AF555 and Hoechst 33342 to visualize cells and nuclei borders. Automated high-content imaging was performed with a spinning-disk microscope to collect 3D image stacks. An artificial intelligence algorithm was trained to segment RPE cells and nuclei borders that were used as a reference to report organelles and structures’ location. Each tagged organelle was segmented with a unique combination of classical and traditional image processing segmentation algorithms. The quality of segmentation was visually inspected by two experts to select the best parameters for the algorithms. Organelle location and morphometry were calculated for each condition to generate reference maps representative of cell states at different timepoints of polarization.

Back to article page