Extended Data Fig. 10: Automatic detection of cell divisions in nematode (Caenorhabditis elegans) embryos.
From: Uncovering developmental time and tempo using deep learning
The Twin Network detects and partitions development into phases that are in line with human staging and early embryogenesis descriptions (http://www.wormbook.org)14. (a) A selected set of images from autostages and boundaries for a single embryo. (b) Calculated similarities used as the basis for the depicted images. Gross homology to a distant morphology at 5–7 min is present around 20–32 min of the acquisition. (c) Dashed lines indicate cytokinesis phases as detected by a human observing the original video. Automatically generated autostages were calculated based on phases of high similarity in embryo morphology and are shown below (frames taken in intervals of 17.5 s). Notably, the blastomere divisions giving rise to ABa, ABp, EMS and P2 cells were correctly identified; n = 1 embryo. Images in (a) correspond to the pictograms in (c) at the indicated timepoints. Scale bar, 10 μm. Raw data from https://www.youtube.com/watch?v=M2ApXHhYbaw.
