MorphoFeatures: unsupervised exploration of cell types, tissues and organs in volume electron microscopy

Abstract

Electron microscopy (EM) provides a uniquely detailed view of cellular morphology, including organelles and fine subcellular ultrastructure. While the acquisition and (semi-)automatic segmentation of multicellular EM volumes is now becoming routine, large-scale analysis remains severely limited by the lack of generally applicable pipelines for automatic extraction of comprehensive morphological descriptors. Here, we present a novel unsupervised method for learning cellular morphology features directly from 3D EM data: a convolutional neural network delivers a representation of cells by shape and ultrastructure. Applied to the full volume of an entire three-segmented worm of the annelid Platynereis dumerilii, it yields a visually consistent grouping of cells supported by specific gene expression profiles. Integration of features across spatial neighbours can retrieve tissues and organs, revealing, for example, a detailed organization of the animal foregut. We envision that the unbiased nature of the proposed morphological descriptors will enable rapid exploration of very different biological questions in large EM volumes, greatly increasing the impact of these invaluable, but costly resources.

Virtual Fly Brain—An interactive atlas of the Drosophila nervous system

“…In the near future, VFB will ingest multiple large connectomics datasets with variable coverage and accuracy of neuron type annotation. BLAST-like algorithms, in the short-term NBLAST for morphology, but longer term supplemented by CBLAST ( Scheffer et al, 2020 ) for connectivity and potentially methods that use subcellular features ( Schubert et al, 2019 ; Zinchenko et al, 2022 ), will be critical to help users to interpret this data by facilitating prediction and assignment of neuron types. For example, a user finding paths between untyped neurons from FlyWire using our circuit browsing tool will be able to use NBLAST to find predicted types for neurons in the circuit, where these exist in other reference data sets.…”

Spatial Transcriptomics-correlated Electron Microscopy

“…The current revolution in the generation of large volume EM datasets calls for new strategies for the annotation of cells. Currently, the cell identities in EM datasets are mostly manually annotated in an unintentionally biased way 21, 22 . Large variations between manual annotations limits the application of deep-learning approaches for automation.…”