Deep learning is emerging as a powerful approach for bioimage analysis. Its use in cell tracking is limited by the scarcity of annotated data for the training of deep-learning models. Moreover, annotation, training, prediction, and proofreading currently lack a unified user interface. We present ELEPHANT, an interactive platform for 3D cell tracking that addresses these challenges by taking an incremental approach to deep learning. ELEPHANT provides an interface that seamlessly integrates cell track annotation, deep learning, prediction, and proofreading. This enables users to implement cycles of incremental learning starting from a few annotated nuclei. Successive prediction-validation cycles enrich the training data, leading to rapid improvements in tracking performance. We test the software’s performance against state-of-the-art methods and track lineages spanning the entire course of leg regeneration in a crustacean over 1 week (504 timepoints). ELEPHANT yields accurate, fully-validated cell lineages with a modest investment in time and effort.
Animals can regenerate complex organs, yet this process frequently results in imprecise replicas of the original structure. In the crustacean Parhyale , embryonic and regenerating legs differ in gene expression dynamics but produce apparently similar mature structures. We examine the fidelity of Parhyale leg regeneration using complementary approaches to investigate microanatomy, sensory function, cellular composition, and cell molecular profiles. We find that regeneration precisely replicates the complex microanatomy and spatial distribution of external sensory organs and restores their sensory function. Single-nuclei sequencing shows that regenerated and uninjured legs are indistinguishable in terms of cell-type composition and transcriptional profiles. This remarkable fidelity highlights the ability of organisms to achieve identical outcomes via distinct processes.
During menstruation, an inner layer of the endometrium is selectively shed, while an outer, progenitor-containing layer is preserved to support repeated regeneration. Progress in understanding this compartmentalization has been hindered by the lack of suitable animal models, as mice and rats do not menstruate. Here, we present transgenic mouse models that recapitulate the key anatomical, functional, and transcriptional features of human menstruation through targeted chemogenetic activation of premenstrual differentiation. Using single-cell spatial transcriptomics, we define a new paradigm for spatially regulated fibroblast differentiation that drives pre-menstrual endometrial layering and ultimately determines the extent of tissue shedding. Our results revise a century-old view of endometrial shedding and regeneration and establish new transgenic mice as powerful tools to advance menstruation research.
Deep learning is emerging as a powerful approach for bioimage analysis. Its use in cell tracking is limited by the scarcity of annotated data for the training of deep learning models. Moreover, annotation, training, prediction and proofreading currently lack a unified user interface. We present ELEPHANT, an interactive platform for 3D cell tracking that addresses these challenges by taking an incremental approach to deep learning. ELEPHANT provides an interface that seamlessly integrates cell track annotation, deep learning, prediction and proofreading. This enables users to implement cycles of incremental learning starting from a few annotated nuclei. Successive prediction-validation cycles enrich the training data, leading to rapid improvements in tracking performance. We test the software’s performance against state-of-the-art methods and track lineages spanning the entire course of leg regeneration in a crustacean over 1 week (504 timepoints). ELEPHANT yields accurate, fully-validated cell lineages with a modest investment in time and effort.
Menstruation and pregnancy disrupt substantial proportions of the uterine lining (endometrium). These breaches impose an immense regenerative burden on the luminal epithelium that lines the uterine cavity, which is proposed to be replenished by cells residing in adjoining epithelial glands. Here, we show that the luminal epithelium and glandular epithelium are maintained by separate progenitor populations during homeostasis, induced menstruation, pregnancy, and postpartum repair in mice. These data challenge the gland-centric model of regeneration during these physiological events, although we find that gland cells can resurface the tissue after chemical ablation. Our data indicate that during menstruation, the luminal epithelium bypasses the need for gland contributions by undergoing extensive expansion and morphogenesis to re-epithelialize stromal surfaces as the tissue breaks down. Analogous morphogenesis occurs during gestational remodeling, revealing luminal epithelial expansion as a unifying mechanism enabling simultaneous stromal disruption and re-epithelialization, which may underlie the endometrium’s remarkable resilience to fibrosis. Graphical
CRISPR/Cas9 technology has revolutionized the production of animal models by reducing experimental timelines, slashing costs and streamlining gene editing, leading to a rapid expansion in the number of unique models for the study of human disease and gene function. However, most non-model animals, many of which are important in cancer and aging research, remain recalcitrant to genome engineering due to our limited understanding of their reproductive biology. Many wild rodents that transmit human diseases remain particularly challenging to engineer due to low pregnancy rates, the lack of external copulatory plugs, and susceptibility to premature termination of pregnancy. Here, we present low-cost activity-based estrous tracking for the efficient generation of timed pregnant and pseudopregnant white-footed mice and extend this tracking method to both lab mice and hamsters. Leveraging this technology, we demonstrate the generation of engineered Peromyscus leucopus , the primary reservoir for Lyme disease-causing bacteria and a putative model organism for studies of aging. These tools have broad implications for biomedical research and ecological engineering. Abstract Figure: