Quantitative image analysis for fluorescence microscopy

Student thesis: Doctoral ThesisDoctor of Philosophy


Localisation microscopy enables imaging beyond the Abbé resolution limit and is based on detecting randomly activated single molecules in a sequence of images. A super-resolution image is then reconstructed using these localisations. Analysis of these points can provide quantitative information about number, shape, and size of features. This thesis presents novel approaches to clustering analysis, identification and characterisation of specific features with a known shape, and confirming the presence of clusters in the sample.

Clustering analysis can be used to both detect clusters, and to measure their size and density. Analysis of localisation microscopy images of biological samples is challenging because clusters are usually small and surrounded by relatively high noise. Here, the Rényi divergence, which can be adjusted to the properties of the data was used. This method provides a more precise measurement of the cluster size, than the commonly used Ripley’s K function.

Biological samples are often highly structured, and the distribution of proteins within these structures is of great interest. In this work a type of adhesive structure called podosomes which consist of an f-actin core surrounded by a protein ring was investigated. Custom written software identified podosome rings in images using a circular model and calculated the relative positions of different ring proteins. This information was used to build a model of podosome ring composition.

The appearance of podosome rings imaged with localisation microscopy depends on the sample preparation and image analysis techniques used, sometimes appearing strongly clustered and sometimes continuous. To attempt to distinguish whether the apparent clusters were due to fluorophore reappearance, a microscopy system was developed to measure fluorescence resonance energy transfer with anisotropy.

The different analysis methods presented in this thesis illustrate the ways in which data analysis and experimental methods can provide a better understanding of a biological system.
Date of Award1 Apr 2017
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorSusan Cox (Supervisor) & Gareth E Jones (Supervisor)

Cite this