Research

MALDI Mass Spectrometry Imaging Sample Preparation LC-MS Data Analysis

Matrix Assisted Laser Desorption Laser Ionisation (MALDI)

MALDI is a mass spectrometry technique for ionising large intact molecules which are fragmented by harder ionisation techniques. Analytes are co-crystallised with an organic matrix which absorbs an irradiating laser beam and transfers the energy whilst preventing delicate molecules from being damaged. For more information see the Wikipedia page or a good textbook.

Mass Spectrometry Imaging

Mass Spectrometry Imaging (MSI) is a hyperspectral imaging technique where spatially resolved mass spectra are acquired at known coordinates across the sample. MSI research broadly follows the three image acquisition stages: sample collection and preparation (sectioning, mounting and other processing); instrument development, optimisation and spectra acquisition; and data interpretation and analysis. My research focusses on the first and last of these areas; the sample preparation and data interpretation.

MALDI MSI Sample Preparation

The sample preparation is a unique challenge for MALDI-MSI, where samples must first undergo a matrix coating processing step. I have worked with various samples types to prepare them for MALDI-MSI, including some difficult to handle tissue types. Each sample type comes with its own specific challenges, but all require a robust method for obtaining thin sections, matrix selection and application which minimises variations in crystallisation and monitoring of the response to laser irradiation. Analyte detection is heavily dependent on all these steps so a full optimisation process to ensure reproducible analysis for each sample type is important.

Liquid Chromatography-Mass Spectrometry

LC-MS couples separation in the liquid phase with molecular detection by mass spectrometry. A key advantage of this approach is that fragmentation studies can be performed during data acquisition leading to identification of the molecules present. LC-MS data from samples collected from discrete spatial locations can be visualised with 'ili which allows the rendering of data in 2D and 3D.

Mass Spectrometry Image Interpretation & Data Processing

One of the challenges presented by recent improvements in mass spectrometry imaging hardware is that much larger images can be collected in a much shorter space of time, with a larger number of analytes in each spectrum. Part of my research focuses on extracting the information contained within the data and applying it to problems of data visualisation, image segmentation and tissue type classification. The approaches I use typically involve subspace projections to reduce the dimensionality of the data followed by both supervised and unsupervised machine learning tools.