(BBSRC DTP) Mapping Notch trafficking using immunofluorescence and computer vision

The functional regulation of proteins in cells is strongly related to their spatial distributions. This is particularly true of signalling proteins, which may be switched on and off by transfer between different locations. Mutations that alter signalling often do so because they change the distributions of proteins. Accurate and detailed mapping of these spatial distributions therefore provides a route to a functional understanding of many aspects of cell biology.
Fluorescence microscopy can visualise individual proteins, allowing their spatial distribution to be mapped in relation to the various organelles and structures within the cell. However, cells are three dimensional, their shapes are heterogeneous, and protein regulation is a dynamic, real-time process involving movement of proteins around the cell. Understanding the regulation of signalling proteins in detail therefore requires the tracking of hundreds, or even thousands, of individual proteins across time-series images, making manual analysis prohibitively time-consuming.
The aim of this project is to develop computer vision algorithms that can automate the process of localising and tracking proteins in time-series fluorescence microscopy images. It will focus on the Notch receptor protein as a model system. Notch is a membrane bound signalling receptor that plays a fundamental role in controlling cell fate, both during development and in the renewal and maintenance of adult tissues and organs. This is an interdisciplinary project and the student will work with experts in both cell biology and computer vision. They will develop the skills required to introduce mutations into fluorescent protein tagged Notch proteins and express in these in cell cultures, and to image these cells using fluorescence microscopy. They will also develop computer vision algorithms that can automatically localise arbitrary numbers of fluorophores in these images, track them across time series, and perform statistical analyses of their spatial distribution in relation to cell structures. They will investigate both classic computer vision algorithms, such as Dirichlet process Gaussian mixture models, and state-of-the-art Deep Learning techniques based on convolutional neural networks. Successful completion will provide the technologies necessary to quantify protein distributions and, in turn, to answer fundamental questions about signalling processes within cells, at levels of detail that would be impossible without the combination of biological and computational techniques.

https://www.research.manchester.ac.uk/portal/martin.baron.html
http://www.stopfrac.com
https://www.tina-vision.net

Entry Requirements:
Applications are invited from UK/EU nationals only. Applicants must have obtained, or be about to obtain, at least an upper second class honours degree (or equivalent) in a relevant subject.