Speedy Construction

When you really want to zoom into a tissue, electron microscopy (EM) is the way to go. Detailed EM images can then be manually 'stitched together', forming 3D reconstructions. For accurate 'stitching' specific features that indicate different cell types must be identified first. This is incredibly time-consuming. Researchers now present an automatic way to create 3D reconstructions using a neural network and deep learning. This approach was tested on EM images of an annelid worm. Different cell types and tissues were identified and reconstructed in 3D, including the foregut, pictured in its entirety (left) and with different muscle groups removed (middle, right). Comparing these results to an atlas of genes across an entire annelid, they found that clusters of distinct cell types identified with this method matched clusters of cells with specific genes switched on. This supports the effectiveness of this speedy approach for analysing large volumes of EM data.

Written by Lux Fatimathas

Lux Fatimathas

Writer

Having pursued science through a career in research, communications and publishing, she is currently a writer and editor for a healthcare marketing agency and also juggles freelance science writing.

• Image from work by Valentyna Zinchenko and Johannes Hugger, and colleagues
• Cell Biology and Biophysics Unit and the Developmental Biology Unit, European Molecular Biology Laboratory (EMBL), Heidelberg, Germany
• Image originally published with a Creative Commons Attribution 4.0 International (CC BY 4.0)
• Published in eLife, February 2023