CLEM-Reg: An automated registration algorithm for correlative light and volume electron microscopy

Session

New and Emerging Concepts in Microscopy

Authors

Daniel Krentzel (2, 1), Matouš Elphick (1, 5), Marie-Charlotte Domart (1), Christopher Peddie (1), Romain Laine (4), Ricardo Henriques (3), Lucy Collinson (1), Martin Jones (1)

Affiliations

1. Francis Crick Institute
2. Institut Pasteur
3. Instituto Gulbenkian de Ciencia
4. UCL
5. Newcastle University

Keywords

• Volume electron microscopy
• CLEM
• vCLEM
• Alignment
• Registration
• Napari

Abstract text

Correlative light and volume electron microscopy (vCLEM) is a powerful imaging technique that enables visualisation of fluorescently labelled proteins within their ultrastructural context on a subcellular level. Currently, expert microscopists find the alignment between datasets by manually placing landmarks on structures that can be recognised in both imaging modalities. However, the manual nature of this process severely impacts throughput and is known to introduce bias. Here, we present CLEM-Reg, a novel algorithm that fully automates the alignment of vCLEM datasets with a hybrid approach combining classical image processing, deep learning and state-of-the-art registration methods, achieving near expert-level registration performance. To ensure easy adoption of CLEM-Reg, we also make a napari plugin available that allows users to run our algorithm end-to-end.