AI-driven processing of Correlative Multimodal Microscopy Images for enhanced analytical insights in disease and therapeutic research

Supervisor Organisation PhD Awarding Entity: phd location
SiriusXT Ltd
University College Dublin
SiriusXT, Dublin, Ireland

Research Focus

The project aims to revolutionize correlative microscopy by seamlessly integrating soft X-ray tomography (cryo-SXT) bridging resolutions and contrast mechanisms of light and electron microscopy. This will be achieved by developing experimental workflows and cutting-edge software capable of registering and segmenting two- and three-dimensional imaging data from multiple microscopy modalities.

Building upon the latest achievements in correlative microscopy, we will adapt current AI-powered image processing and quality assessment algorithms and develop new ones. This project involves hands-on experiments with other project CLEXM Partners, who operate leading disease research microscopy labs, to establish sample-handling workflows and conduct ground-breaking experiments to push the boundaries of scientific knowledge in biomedicine.


Correlative imaging data of biological specimens collected by CLEXM partners using cryo-SXT, cryo-ET, cryo-FIB/SEM, cryo-FM, cryo-3D-SIM, and hard X-ray microscopy will be shared with SiriusXT for their correlative integration and the development of segmentation workflows.

A survey of state-of-the-art correlative data registration and segmentation algorithms will be made. Algorithms and workflows for further development and incorporation into CLEXM software will be selected based on their performance and suitability. Sample preparation, handling protocols, correlative registration, and segmentation software development will be done in collaboration with the consortium partners.

Aim 1

Image Integration and analysis

Develop approaches to integrate data across imaging modalities and resolution scales building upon best practices used in electron tomography and fluorescence microscopy fields

Implement current or develop new multimodal 3D data segmentation algorithms to aid biological interpretation

Aim 2

Sample preparation and handling

Establish sample preparation and handling protocols to aid correlative data collection and integration

Aim 3


– Apply the developed software and sample handling protocols to aid biological interpretation in partner research

Pictures Attached

Fig. 1 (Left) Fluorescence image of a HeLa cell. (Right) Soft X-ray tomogram of the imaged cell with a 3D-segmented mitochondria shown in yellow.


Role/Focus of PhD:

The PhD Doctoral Network candidate (DNC) will be responsible for developing an in-depth understanding of image processing and analysis workflows. The DNC will select the best image registration and semantic segmentation workflow algorithms based on the ease of implementation, such as training of a neural network for X-ray datasets, the effort necessary to segment a dataset, and the accuracy of segmentation, for which a suitable metric will be decided in the course of the project.

The PhD candidate should fulfil the following requirements:

  1. Educational Background:

Minimum: BSc in Computer Science or Bioinformatics (or equivalent) with experience in AI and other image analysis techniques

Ideally: A Master’s degree (or equivalent) in Bioinformatics or Computer Science. Knowledge of microscopy will be considered an advantage.

  1. Research Skills:

– Knowledge of image acquisition and data handling in microscopy.

– Experience in preparation and handling biological specimens for imaging, particularly for soft X-ray cryo-tomography or cryo-electron microscopy, will be considered an advantage.

  1. Programming and Software Skills:

– Proficiency in Python

– Experience in developing and implementing software algorithms for image processing and analysis is an advantage.

  1. Communication and Collaboration:

– Strong written and verbal communication skills for effectively documenting research findings and presenting results to scientific audiences.

– Ability to work both independently and collaboratively as part of a research team.

 – Willingness to learn and adapt to new technologies and methodologies.

  1. Additional Skills (Desirable but not essential):

– Knowledge of X-ray imaging and tomography techniques.

– Familiarity with deep learning frameworks (e.g., TensorFlow, PyTorch) for developing and training neural networks and applying them to image analysis tasks.

– Knowledge of image registration techniques and algorithms.

– Understanding of computer vision concepts and techniques for object classification and semantic segmentation.

– Experience in scientific writing, including preparing manuscripts and conference presentations.


Application Deadline: 18th August 2023 (Closed)