The Microscopy Bioscience Platform offers comprehensive advanced image analysis support tailored to the diverse needs of researchers.
For projects requiring robust, quantitative analysis in biology, our dedicated image analysis team works within the Technology Innovation pillar, in the Cambridge Advanced Imaging Centre and integrates computational methods with modern microscopy. This team supports cutting-edge, collaborative projects, driving innovation in image analysis methodologies and tools. The team develops custom workflows and novel algorithms, addressing challenges such as processing large-scale images and handling complex biological data to achieve high quality, reproducible research. Examples of our work are provided below
We also provide access to dedicated workstations, equipped with state-of the art software (Imaris, Huygens, Arivis). For routine analysis, our Light Microscopy team and image analysis specialists provide advice and suggestions for resources.
Research Technical Professionals in our Platforms and Facilities provide essential scientific support, and we follow the Royal Microscopy Society guidelines for fair attribution, ensuring their contributions are routinely acknowledged in publications, with authorship considered for substantial intellectual input. See our Fair Attribution Policy on the Platform overview page for more information.
Contact
If you would like support for advanced image analysis in your project, please contact Dr Leila Muresan.
Image Analysis Team
Dr Leila Muresan
Senior Research Associate/
Research Software Engineer
lam94 [at] cam [dot] ac [dot] uk
Training and Drop-in clinics
Image analysis clinic:
Mondays 1pm - 2:30 pm (in the Wolfson Imaging Centre, Anatomy building basement, Downing site)
Please attend the Light Microscopy clinics for advice on routine image processing tasks.
Training courses (once per term):
We will be providing training modules mixing theoretical and problem-solving based approaches for image analysis:
-
Image processing with Fiji
-
Image processing with Python
-
Image processing with Matlab
These are introductory courses, programming skills are welcome but not mandatory. You will learn what is an image, how to open, save, filter, segment a (multi-dimensional) image, plus a few microscopy specific features.
Commercial Software Provision
Image analysis workstations hosting commercial image analysis software are provided in the Anatomy building. Available software includes:
- Imaris 10
- Zeiss Arivis Vision4D
- Zeiss Zen 3.10
To book time on the image analysis workstations, please use PPMS.
Image processing workshops:
- Matlab tutorial (small groups, upon request)
- Twice termly drop-in clinics
Advanced Image Analysis Projects
Cell tracking and segmentation in 3D timelapse images
A deep learning–based algorithm was developed to segment cell nuclei in anisotropic 3D light-sheet microscopy images of pre-implantation embryos. In addition, a novel tracking method was implemented to monitor nuclear dynamics over a 24-hour period. To date, more than 15 embryos have been successfully processed, with additional datasets currently underway. Furthermore, 3D ground truth training data were generated through an automated pipeline and subsequently refined via manual curation.
Collaboration with the Niakan group (PDN).
Reference: Ahmed Abdelbaki, Afshan McCarthy, Anita Karsa, Leila Muresan, Kay Elder, Athanasios Papathanasiou, Phil Snell, Leila Christie, Martin Wilding, Benjamin J. Steventon, Kathy K. Niakan - Live imaging human embryos reveals mitotic errors and lineage specification prior to implantation. bioRxiv 2024.09.26.614906; doi: https://doi.org/10.1101/2024.09.26.614906
Denoising for large scale images
We explored scalable approaches to pre-processing large microscopy datasets—such as those generated by light-sheet microscopy—using DASK, a general-purpose library for parallel computing. As a representative application, we focused on 3D + time denoising. By leveraging the capabilities of high-performance computing (HPC), our results demonstrate improved performance and scalability, exceeding the quality and efficiency of previously reported methods.
Collaboration with Jerome Boulanger, (MRC-LMB), Simon Clifford (UIS) and Clare Buckley (University of Manchester).
Light sheet microscopy image deconvolution
We developed a novel model for light-sheet image formation that incorporates a spatially varying point spread function (PSF), enabling more accurate and adaptive deconvolution of light-sheet microscopy images. The deconvolution is formulated as a variational optimization problem and solved using a primal-dual algorithm, explicitly accounting for realistic Poisson-Gaussian noise. To further improve performance and generalizability, we implemented automated regularization parameter tuning and PSF fitting, including the correction of optical aberrations.
Collaboration with Jerome Boulanger, (MRC-LMB), Yury Korolev (Bath), and Carola Schoenlieb (DAMTP).
Reference: Toader B, Boulanger J, Korolev Y. et al. Image Reconstruction in Light-Sheet Microscopy: Spatially Varying Deconvolution and Mixed Noise. J Math Imaging Vis 64, 968–992 (2022). https://doi.org/10.1007/s10851-022-01100-3
Single molecule localisation microscopy
Single particle tracking (Source)
Pipelines were developed to compute and analyse single-molecule trajectories under challenging signal-to-noise conditions. Quantifying the statistical properties of transcription factor trajectories within the nuclei of fruit fly salivary gland cells, allowed gaining deeper insights into the mechanisms underlying gene regulation.
Collaboration with the Bray group (Department of Physiology, Development and Neuroscience).
References:
Baloul S, Roussos C, Gomez-Lamarca M, Muresan L, Bray S. Changes in searching behaviour of CSL transcription complexes in Notch active conditions. Life Sci Alliance. 2023 Dec 14;7(3):e202302336. doi: 10.26508/lsa.202302336. PMID: 38097371; PMCID: PMC10721712.
Gomez-Lamarca MJ, Falo-Sanjuan J, Stojnic R, Abdul Rehman S, Muresan L, Jones ML, Pillidge Z, Cerda-Moya G, Yuan Z, Baloul S, Valenti P, Bystricky K, Payre F, O'Holleran K, Kovall R, Bray SJ. Activation of the Notch Signaling Pathway In Vivo Elicits Changes in CSL Nuclear Dynamics. Dev Cell. 2018 Mar 12;44(5):611-623.e7. doi: 10.1016/j.devcel.2018.01.020.
Clustering
Spatial statistics approaches were developed, implemented, and applied to study chromatin loop structures imaged using single-molecule localization microscopy (STORM). Key challenges—such as testing for molecular clustering, estimating cluster sizes, and assessing mutual exclusion between different molecule types in dual-colour images—were successfully addressed.
Collaboration with the White group (PDN).
References:
Ball ML, Koestler SA, Muresan L, Rehman SA, O’Holleran K, et al. (2023) The anatomy of transcriptionally active chromatin loops in Drosophila primary spermatocytes using super-resolution microscopy. PLOS Genetics 19(3): e1010654. https://doi.org/10.1371/journal.pgen.1010654
Koestler SA, Ball ML, Muresan L, Dinakaran V, White R. Transcriptionally active chromatin loops contain both 'active' and 'inactive' histone modifications that exhibit exclusivity at the level of nucleosome clusters. Epigenetics Chromatin. 2024 Mar 25;17(1):8. doi: 10.1186/s13072-024-00535-9. PMID: 38528624; PMCID: PMC10962081.
Other services available in the School of Biological Sciences
GIIF – Gurdon Institute Imaging Facility Website
The Gurdon Institute provides dedicated image analysis facilities and support as part of the wider imaging facility and comprehensive training.
