Deep learning is widely applicable to phenotyping embryonic development and disease.

Genome editing simplifies the generation of new animal models for congenital disorders. However, the detailed and unbiased phenotypic assessment of altered embryonic development remains a challenge. Here, we explore how deep learning (U-Net) can automate segmentation tasks in various imaging modalities, and we quantify phenotypes of altered renal, neural and craniofacial development in Xenopus embryos in comparison with normal variability. We demonstrate the utility of this approach in embryos with polycystic kidneys (pkd1 and pkd2) and craniofacial dysmorphia (six1). We highlight how in toto light-sheet microscopy facilitates accurate reconstruction of brain and craniofacial structures within X. tropicalis embryos upon dyrk1a and six1 loss of function or treatment with retinoic acid inhibitors. These tools increase the sensitivity and throughput of evaluating developmental malformations caused by chemical or genetic disruption. Furthermore, we provide a library of pre-trained networks and detailed instructions for applying deep learning to the reader's own datasets. We demonstrate the versatility, precision and scalability of deep neural network phenotyping on embryonic disease models. By combining light-sheet microscopy and deep learning, we provide a framework for higher-throughput characterization of embryonic model organisms. This article has an associated 'The people behind the papers' interview.

Parallel in vivo analysis of large-effect autism genes implicates cortical neurogenesis and estrogen in risk and resilience.

Gene Ontology analyses of autism spectrum disorders (ASD) risk genes have repeatedly highlighted synaptic function and transcriptional regulation as key points of convergence. However, these analyses rely on incomplete knowledge of gene function across brain development. Here we leverage Xenopus tropicalis to study in vivo ten genes with the strongest statistical evidence for association with ASD. All genes are expressed in developing telencephalon at time points mapping to human mid-prenatal development, and mutations lead to an increase in the ratio of neural progenitor cells to maturing neurons, supporting previous in silico systems biological findings implicating cortical neurons in ASD vulnerability, but expanding the range of convergent functions to include neurogenesis. Systematic chemical screening identifies that estrogen, via Sonic hedgehog signaling, rescues this convergent phenotype in Xenopus and human models of brain development, suggesting a resilience factor that may mitigate a range of ASD genetic risks.

Localized JNK signaling regulates organ size during development.

A fundamental question of biology is what determines organ size. Despite demonstrations that factors within organs determine their sizes, intrinsic size control mechanisms remain elusive. Here we show that Drosophila wing size is regulated by JNK signaling during development. JNK is active in a stripe along the center of developing wings, and modulating JNK signaling within this stripe changes organ size. This JNK stripe influences proliferation in a non-canonical, Jun-independent manner by inhibiting the Hippo pathway. Localized JNK activity is established by Hedgehog signaling, where Ci elevates dTRAF1 expression. As the dTRAF1 homolog, TRAF4, is amplified in numerous cancers, these findings provide a new mechanism for how the Hedgehog pathway could contribute to tumorigenesis, and, more importantly, provides a new strategy for cancer therapies. Finally, modulation of JNK signaling centers in developing antennae and legs changes their sizes, suggesting a more generalizable role for JNK signaling in developmental organ size control.