Poster Presentation 24th Annual Lorne Proteomics Symposium 2019

A multidisciplinary approach to the analysis of complex MSI data (#91)

Matthew O'Rourke 1 , Ben Roediger 2 , Matt Padula 3
  1. The University of Sydney, St Leonards, NSW, Australia
  2. Centenary institute, Sydney
  3. University of Technology Sydney, Sydney

Mass Spectrometry Imaging (MSI) is a hybrid technique combining the visualisation features of traditional histology with accompanying  mass spectrometric data. MSI uses a Matrix Assisted Laser Desorption/ Ionisation (MALDI) mass spectrometer to create spatial maps of targeted biomolecule classes. The spatial maps take the form of 2D images that show the location and relative intensity of a given molecule and the dynamic range of the instrument allows for analyses of virtually any class of ionisable biomolecule without the need for adding detection reporter probes. Though this may seem straightforward the issue becomes the translation of mass spectrometric data; a complex list of numbers, to literal pictures, and how these pictures need to be interpreted. When MSI data is viewed from the perspective of microscopy, MSI data is low resolution, pixelated and does not allow the visualisation or distinction of the whole tissue. By all these counts, the data is not of a publishable quality. Conversely from a mass spectrometry perspective, the mass resolution is very low and the number of detected/ detectible molecules and the confidence that molecular identities are statistically confident is tenuous.

We therefore describe a new approach to the analysis of MSI data that converts it to 2 dimensional images, and then processes them using an Image J based processing pipeline.  The pipeline is then able to report R2 values for the correlation of the detected peptides, thereby confirming their co-localisation on tissue. By using a pixel to pixel correlation of both intensity and presence and absence of a series of peptides for a single protein, peptides can be matched together based on location, to form peptide mass fingerprints (PMF’s) which can then be searched through traditional PMF databases to achieve reliable protein identifications.  Therefore, this approach allows MSI data to be analysed in isolation without requiring complementation with other instrumentation such as LCMS.