FindAPhD LIVE! Study Fair

Edinburgh | Oxford | Leeds

FindA University Ltd Featured PhD Programmes
Norwich Research Park Featured PhD Programmes
Imperial College London Featured PhD Programmes
National University of Ireland, Galway Featured PhD Programmes
University of Warwick Featured PhD Programmes

Applying Machine Learning to 3D single molecule localisation analysis in super-resolution microscopy

This project is no longer listed in the FindAPhD
database and may not be available.

Click here to search the FindAPhD database
for PhD studentship opportunities
  • Full or part time
    Prof M Peckham
  • Application Deadline
    No more applications being accepted
  • Competition Funded PhD Project (European/UK Students Only)
    Competition Funded PhD Project (European/UK Students Only)

Project Description

with 2nd Supervisor Dr Joanna Leng, in the School of Computing.
Photoactivated localisation microscopy (PALM) and direct stochastic optical reconstruction microscopy (dSTORM) are two single molecule approaches that allow us to interrogate the organisation of specific proteins within a cell to very high resolution (a few nm) using light microscopy. PALM uses genetically encoded fluorescent proteins, while dSTORM uses fluorescent dye labels for specific proteins (e.g. dye labelled antibodies or other non-antibody binding proteins such as Affimers).
At Leeds, we use an astigmatic lens based approach to generate a 3D PALM or dSTORM image (e.g. see Tiede et al., 2017, Lambacher et al., 2016). To process the image data and generate a 3D reconstruction image, each fluorescent spot is compared to a reference dataset, which we obtain separately using fiducial markers, to enable us to accurately position each spot accurately in x,y and z (lateral and axial planes) as described (York et al., Nature Methods, 2011). In addition, this approach can ‘link’ fluorescent spots, where the same fluorescent particle appears in several frames (same x,y,z co-ordinates) before it stops fluorescing. This is in effect a Machine Learning process, as each individual image is compared to a reference image to determine where the fluorescent spot is in x,y and z. However, this is a lengthy process that currently cannot be done in real time and requires at least 12 hours to process and image.
The student will work on developing and improving this approach to a) adapt it to work on a GPU based interface with use of the local high-performance computing resources, which are managed by the ARC (Advanced Research Computing) team and that are free at the point of use for all researchers at the University of Leeds, with the aim of generating 3D reconstructions of super-resolution images as rapidly as possible, potentially in ‘real time’. As part of this approach, they will also interrogate and improve on our approach to ‘linking’ particles between adjacent frames, to better address issues with overlapping particles, with the potential of developing and using a Bayesian approach in this analysis, enabling us to capture more information and generate more detailed 3D super-resolution maps, and this would implement a Machine Learning base strategy. We already have multiple datasets that we can use for this project, including cytoskeletal proteins, muscle Z-discs, and virus replication centres. The overall aim is to generate fast detailed ‘super-resolved’ 3D images of protein organisation in cells, which fits within BBSRC remit across a number of different areas.

Funding Notes

Project is eligible for funding under the BBSRC White Rose DTP: Doctoral Studentships in Artificial Intelligence, Machine Learning and Data Driven Economy.

Successful candidates will receive funding for 4 years, covering UK/EU fees and research council stipend (2018-19: £14,777).

Candidates should have/be expecting a 2.1 or above at undergraduate level in a relevant field. If English is not your first language, you must also meet our language entry requirements. EU candidates are subject to RCUK residency requirements. The PhD is to start in Oct 2018.

Apply online: https://studentservices.leeds.ac.uk/pls/banprod/bwskalog_uol.P_DispLoginNon

Include project title and supervisor name, and upload a CV and transcripts.

References

Tiede et al., 2017, ELife 'Affimer proteins are versatile and renewable affinity reagents'
Lambacher et al., 2016 Nature Cell Biology 'TMEM107 recruits ciliopathy proteins to subdomains of the ciliary transition zone and causes Joubert syndrome'


How good is research at University of Leeds in Biological Sciences?

FTE Category A staff submitted: 60.90

Research output data provided by the Research Excellence Framework (REF)

Click here to see the results for all UK universities

Let us know you agree to cookies

We use cookies to give you the best online experience. By continuing, we'll assume that you're happy to receive all cookies on this website. To read our privacy policy click here

Ok