When
Where
Speaker: Woody March-Steinman, Applied Mathematics
Title: Cryptid Stories for October
Abstract: Suppose you have a colonoscopy, and agree to donate a tissue sample to science. What can be done with your generosity? In some cases, these samples can be grown indefinitely on a dish. They can then be studied for a better understanding of the signals that influence tissue patterning and heterogeneity. This emerging experimental model requires new tools to identify and analyze important cell and tissue morphologies under drug perturbations. I aim to discuss the development of NodeNet as a pipeline for evaluating tissue regions for stem-ness (crypts) and integrating biologically-relevant loss functions for performance under human labeling uncertainty. I’ll also show the utility of this analysis toolkit for hypothesis generation from a large dataset of organoid movies.