Investigation and interpretation of large mass spectrometry imaging datasets

Race, Alan M. (2016). Investigation and interpretation of large mass spectrometry imaging datasets. University of Birmingham. Ph.D.

Preview

Race16PhD.pdf
PDF - Accepted Version
Download (5MB)

Abstract

Mass spectrometry imaging (MSI) enables two- and three-dimensional overviews of hundreds of unlabelled molecular species including drugs, metabolites, lipids and proteins in complex samples such as intact tissue. In this research, a new extensible software platform is presented, suitable for spectral preprocessing, multivariate analysis and visualisation of large MSI datasets from all major MSI vendors.
Principal component analysis (PCA) has been widely used in the unsupervised processing of MSI data. Standard implementations of PCA require the entire dataset to be stored in memory, necessitating a compromise between the number of pixels and the number of peaks to include. In this research a new method which has no limitation on the number of pixels is developed. Hierarchical composition of data has been shown as an efficient method of capturing the information present within images in other fields. An adaptation of these ideas to MSI data is described. The way in which imaging data are presented can have a significant impact on the perceived structure, especially when using false colour to display images. The research presented in this thesis has resulted in new recommendations for presentation of MS images. Finally, the software and algorithms presented were used to analyse MSI data from a traumatic brain injury model. Manual exploration and use of multivariate analysis methods such as PCA did not reveal any differences between the injured hemisphere of the brain and the control hemisphere, however the hierarchical composition algorithms identified multiple ion images which appear elevated in the injured hemisphere.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):