Optimising high resolution measurements of epidermal growth factor receptor oligomers in cells using Machine Learning algorithms

Student thesis: Doctoral ThesisDoctor of Philosophy


The epidermal growth factor receptor (EGFR) is a cell surface receptor, which controls cell growth and division. Mutations affecting the receptor expression could lead to cancer. Analysis of EGFR interactions with living cells requires measuring separations between 5 and 60nm. The separations are calculated by analysing time-series of diffraction limited spots, generated by labelled EGFRs. Finding such time-series manually is time consuming and non-reproducible. This project uses machine learning algorithms in combination with understanding of the data collection process and analysis requirements to optimise the data selection process, by automatically rejecting non-analysable time-series. The comparison to the manual process shows that the automated process significantly decreases the time required for data selection and decreases the uncertainty in the distance measurements.
In the last chapter of the thesis the refined data selection process is used as part of the analysis of EGFR experiments, which studies the effects of phorbol myristate acetate (PMA) on the receptor dimerisation. The results from the experiment show consistencies with data from electron microscopy and also that the PMA drives the EGFR to form smaller clusters as compared to the control experiment, however there is large uncertainty in the result due to limited data set.
Date of Award1 Oct 2021
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorSimon Ameer-Beg (Supervisor) & Marisa L. Martin-Fernandez (Supervisor)

Cite this