10th-11th October 2016

The workshop takes place at the Radcliffe Conference Centre

Organisers: Dr Magnus Richardson and Dr Till Bretschneider

This small workshop will bring together researchers using a variety of imaging and image-analysis techniques at the cellular and tissue levels.

Monday 10th October

1.00pm Registration and Welcome
1.30pm Dr Guy Blanchard, University of Cambridge
Using image analysis to infer epithelial mechanics in vivo
2.15pm Dr Boguslaw Obara, University of Durham
BioImage Informatics: Where BioImaging Meets Computer Science
3.00pm Dr Masanori Mishima, University of Warwick
Label-free automated tracking of cell division dynamics in C. elegans embryos
3.45pm-4.15pm coffee + poster session
4.15pm Prof Gail McConnell, University of Strathclyde
The Mesolens: a new microscope objective for imaging large biomedical specimens with sub-cellular 3D resolution
5.00pm Dr Derek Magee, University of Leeds
Histopathology Image Analysis on a Large Scale

Tuesday 11th October

9.00am Dr Nasir Rajpoot, University of Warwick
Morphometric Profiling in Digital Pathology
9.45am Dr David Gurevich, University of Bristol
The role of macrophages and inflammation in wound neoangiogenesis
10.30am-11am coffee + poster session
11.00am Dr Jason Berwick, University of Sheffield
Does the fMRI BOLD signal always reflect changes in neuronal activity?
11.45am Dr Karuna Sampath, University of Warwick
Functional analysis of RNA elements by live imaging in zebrafish embryos
12.30pm Close

Due to the limited attendance would we like to encourage those that register to consider presenting a poster. If you will be presenting a poster, please contact Dr Magnus Richardson (magnus.richardson@warwick.ac.uk) with the title and abstract.

Abstracts

Dr Guy Blanchard, University of Cambridge: Using image analysis to infer epithelial mechanics in vivo

Recent advances in the live imaging of embryonic development promise to revolutionise our mechanistic understanding of morphogenesis. Tissue mechanics are a vital link between cellular protein expression and the changing shapes of embryos, but remain probably the least well understood aspect of development. A biomechanical understanding requires that four hurdles be overcome. First, 4D imaging datasets need to be translated into cell trajectories and cell shapes over time. Second, local tissue deformation must be quantified and broken down into the additive contributions of local cell behaviours. Third, the fluorescence intensity of Myosin motors can be quantified as a proxy for cell contractility. The final step is to combine the above information to estimate mechanical parameters in vivo. I will summarise methods we are using to achieve these four steps in multiple tissues of the Drosophila fly embryo. We find that supra-cellular mechanical structures are important drivers of tissue morphogenesis.

Dr Boguslaw Obara, University of Durham: BioImage Informatics: Where BioImaging Meets Computer Science

Recent advances in imaging technologies are now enabling us to acquire image data, where the sheer size and complexity of the data exceeds the capacity of human analysis, creating a fundamental barrier that hinders the potential impact of imaging-enabled biology. New revolutions in the field can be made possible only by developing the tools, technologies and workflows to extract and exploit information from imaging data so as to achieve new fundamental biological insights and understanding, as well as developing possible strategies in biotechnological applications. Here, we propose to integrate bioimaging data and bioimage informatics to investigate the multi-scale topology and dynamics of complex biological systems. We have developed a wide range of high-throughput image processing approaches specifically designed to enhance, extract, analyse and model a wide range of biological structures, including molecules, plant cells and their organelles, pollen tubes, fungal and cytoskeletal networks, leaves, and many more, from multidimensional images obtained by a wide spectrum of bioimaging modalities.

Dr Masanori Mishima, University of Warwick: Label-free automated tracking of cell division dynamics in C. elegans embryos

Mitotic spindle plays central roles in cell division. Although probes based on fluorescent proteins such as GFP-tubulin allow us to monitor the dynamics of the mitotic spindle in a highly specific manner, there are limitations due to its invasiveness and requirement of genetic manipulations. In contrast, conventional light microscopy can provide us with a broad-range of information in a higher time resolution and in an almost non-invasive manner without genetic manipulation. For example, a trained human eye can easily detect mitotic spindles and chromosomes in images acquired by differential interference contrast (DIC) microscopy. However, automation of image analysis of the DIC movies remains challenging due to low contrast and specify. In this presentation, I will discuss our efforts towards automated tracking of spindle poles and detection of chromosomes in dividing C. elegans embryos using computational methods such as active contour segmentation and deep learning.

Prof Gail McConnell, University of Strathclyde: The Mesolens: a new microscope objective for imaging large biomedical specimens with sub-cellular 3D resolution

Optical lenses reached the limit of resolution set by the wavelength of light more than a century ago. However, no attempt was made to achieve the maximum resolution in the case of low-magnification lenses, probably because the visual image would then have contained detail too fine to be perceived by the human eye. Currently available lenses of less than 10x magnification are of simple construction and their numerical apertures (which determine their resolving power) are 0.2 or less, as compared with 1.3 or more in high-power lenses. They are perfectly adequate for the eye or a standard camera, but in the 1980s, confocal microscopy and improvements in camera resolution revealed a need for better low-power lenses: many researchers found that thin confocal optical sections could not be obtained at 4x magnification because of the poor axial response of these lenses. We have developed a novel lens system called the Mesolens, which achieves an N.A. of nearly 0.5 at a magnification of only 4x. When compared with a standard 4x objective, its lateral resolution is 2.5 to 5 times better and its depth resolution (which is vital for confocal or multi-photon microscopy) is 10x better. This unusually large objective lens provides, for the first time, good optical sectioning of specimens as large as entire 12.5-day mouse embryos (>5 mm long) with sub-cellular detail in every developing organ. We are creating at the University of Strathclyde a facility for the development and application of Mesolenses in biomedical science. Progress in the creation of the facility will be presented, together with recent super-wide-field and confocal mesoscopy datasets.

Dr Derek Magee, University of Leeds

Bio-Medical image analysis has traditionally dealt with images on a rather small scale. Typical images have been limited to a few thousand pixels square. With the advent of Digital pathology much larger images have become available (“virtual slides” can be up to around 100K x 100K pixels). This has led to the development of algorithms and frameworks capable of processing such large images in a number of groups around the world often using High Performance Computing and GPU computing. The University of Leeds has moved one stage beyond this and has been investigating the analysis of aligned stacks of hundreds of such images which form terrapixel 3D data sets. In this talk I will outline our efforts towards full-resolution whole slide analysis of such stacks by efficient and parallelisable analysis methods for cell detection/classification, colour, and tissue type analysis. These utilise highly efficient cell detection and classification algorithms, and deep learning based tissue identification methods that can process whole slides in minutes at full resolution on fairly modest hardware. This facilitates the stacking up of such results as 3D result sets and the microscopic investigation of tissue in a completely new way.

Dr David Gurevich, University of Bristol: The role of macrophages and inflammation in wound neoangiogenesis

Neoangiogenesis is considered an integral part of the wound healing process and many pathologies of repair are associated with impaired wound angiogenesis; however, the mechanisms used to regulate this process during repair are poorly understood. Macrophages play a central in wound healing, and have been shown to contribute to blood vessel recruitment during development and some wound and pathology contexts. How communication proceeds between macrophages and endothelial cells during wound healing remains unclear. We therefore aim to characterise the communications and mechanisms underlying these cellular interactions during wound repair and subsequent resolution. In this study, we investigate the cellular interactions and communications between innate immune cells and endothelial cells at the wound site in real time and in vivo. Neutrophils appear to be peripherally involved in this process, showing little direct contact with or effect on wound blood vessels. By contrast, macrophages become intimately associated with wound blood vessels from very shortly after injury and throughout the repair process. Macrophage ablation results in impaired neoangiogenesis, proportional to the extent of ablation. Manipulation of the wound environment affects macrophage gene expression and subsequent blood vessel recruitment, with premature dampening of the typical inflammatory response resulting in impaired neoangiogenesis. A specific role for macrophages in controlling neoangiogenesis appears to be tied to their ability to detect hypoxia and subsequently transduce this hypoxic signal to the endothelial cells in order to trigger sprouting. Finally, macrophages also appear to play a role in pruning of vessels during the remodelling phase of wound repair, and their absence impairs appropriate vessel pruning post repair.

Dr Jason Berwick, University of Sheffield: Does the fMRI BOLD signal always reflect changes in neuronal activity?

Functional magnetic resonance imaging (fMRI) has revolutionised the field of cognitive neuroscience by increasing the ability to probe the workings of the human brain. However, it does suffer from a limitation that it is only measuring a secondary hemodynamic marker of neural activity termed the blood oxygenation level dependent (BOLD) response. A complete understanding of the BOLD signal source with regard to metabolism and neural activation is still lacking and is of critical importance especially as it is being used as a biomarker in many disease states such as dementia. There is a distinct possibility that BOLD signals in disease may not be driven entirely by neural activation but result as a consequence of a perturbed vascular system. This has implications not only for task evoked BOLD studies but spontaneous connectivity research. I will explore how non-neuronally driven BOLD signals could be generated and show some initial results from pilot experiments aimed at developing a mouse model for neurovascular research.

See also:
Mathematics Research Centre
Mathematical Interdisciplinary Research at Warwick (MIR@W)
Past Events 
Past Symposia