Stereo discovery v8 microscope

Bright-field images were acquired using a Zeiss Stereo Discovery V8 microscope equipped with an AxioCam MRc camera and were processed with AxioVision software (Zeiss). Fluorescence images were acquired using a Zeiss LSM 780 confocal microscope. All images used for cell counting were acquired using a Marianas spinning disk confocal microscope or a Zeiss LSM 780 confocal microscope with a 40× objective. Cells were counted using Imaris 9.1 (Bitplane) or the machine learning–based image segmentation method StarDist (Schmidt et al., 2018 (link)). In the latter case, StarDist was used to detect individual cells in multi-channel images and the average fluorescence intensity of each cell was quantified for each channel. The detection and quantification steps were implemented as a batch execution script in the OMERO server platform (Allan et al., 2012 (link)). The OMERO viewer was used to interactively select the intensity threshold in each channel for cells positive for the marker. For each genotype, 3–6 embryos with 1–5 sections per embryo at comparable positions were quantified. For quantifying cell death, TUNEL positive cells were counted manually. For each genotype, 4–6 embryos with 3 sections of each rostrocaudal position per embryo were quantified. Data points in all quantification graphs correspond to individual embryos.