Columbus software 2

Multiparametric image analysis was performed using Columbus Software 2.8.0 (PerkinElmer). Hoechst signal was used to detect all cell nuclei. Phalloidin-TRITC was used to determine the cytoplasmic region to the corresponding nucleus. Moreover, we applied a filter to remove border objects (nuclei that cross image borders) and cells with extremely small nuclei (dead cells). In a next step, we calculated the morphology and Alexa488 fluorescence intensity in each cell region (nucleus and cytoplasm). In addition, we performed spot detection in the cytoplasm and used morphology and intensity for each spot to define “big spots.” Each spot was detected as a small region within the corresponding image by having a higher intensity than its surrounding area. Furthermore, we selected cells with three or more “big spots” in the cytoplasm and calculated the percentage of “positive” cells in each well. Finally, a hit was defined if the ratio of cytoplasmic intensity and/or the ratio of cells with more than 3 spots was > 1 in at least 3 of 5 plates after RSL3 treatment. An illustration on the automated detection method using the Hoechst, phalloidin-TRITC- and Alexa488 antibody- signal is presented in Figure S1.