Abstract:   
Proteome analysis or proteomics is the direct qualitative and quantitative analysis of the proteine complement in an organism, tissue, or cell under a set of known physiological and environmental conditions. Mass spectrometry, aided by fractionation techniques such as liquid chromatography and gel electrophoresis, has become the central tool for proteomics research. It has been at the basis of a considerable set of successes in fields such as experimental biochemistry and the elucidation of pathologies such as Alzheimer's Disease.

Most proteomics studies, however, have focused solely on the identification and the quantitative analysis of proteins in a sample, without taking the absolute spatial location or exact origin of the sample within tissue into account. The few studies where identified proteins were linked to their spatial origin in a tissue section, demonstrated that this type of information could provide vital clues in gathering further insight into biological processes.

In our work we use the combination of mass spectrometry and spatial information to acquire some insight into the more spatially-oriented biomolecular processes, such as protein distribution/location, and migration. The technology used to gather this type of data is known as laser-based Imaging Mass Spectrometry (IMS). IMS is a relatively new technology that uses the molecular specificity and sensitivity of normal mass spectrometry to collect a direct spatial mapping of biomolecules in given tissue sections. The wet-lab procedure comes down to preparing a tissue section, fixing the section onto a MALDI target plate, applying an appropriate chemical matrix solution, and performing a MALDI mass spectral measurement on each grid point of a virtual array that has been superimposed on the tissue section. The result is an array of 'pixels' covering the tissue section, with a mass spectrum linked to each individual pixel. IMS has been successfully used in a set of pioneering studies that have mainly focused on mammalian tissue.

The main objective of our work is the extension of the in silico analysis of these types of experiments. Essentially, the data that comes out of an IMS experiment has the form of a mathematical space that has two spatial dimensions and as many feature dimensions as the mass range settings of the mass spectrometer allow. Typically, we have initially a grid of pixels in which each element encodes a peaklisted mass spectrum as a vector of intensities for a discretized mass-over-charge value. We will be looking at methods to extract fragment images, protein images, and perform segmentation on these false color images.

The validation and verification of the developed in silico analysis algorithms will be done on the basis of biological/medical case studies.