Use of imaging multivariate analysis to improve biochemical and anatomical discrimination in desorption electrospray ionisation mass spectrometry imaging

Abstract

Desorption electrospray ionisation (DESI) mass spectrometry images usually contain a large amount of information that can be difficult to interpret in an objective manner. We explore the use of imaging multivariate analysis (MVA) on DESI images of protein spots and rat brain sections to automatically assign peaks and improve discrimination of spatially important features. DESI parameters were optimised on an ion trap mass spectrometer for (a) consistent imaging of dried single and mixture spots of insulin, myoglobin and BSA from a Permanox slide, and (b) to produce a MS image of rat brain coronal section at 100 μm resolution. Multivariate curve resolution (MCR), an imaging MVA technique was applied to these images after appropriate data binning. MCR analysis on DESI images of protein mixture spots allowed the multiply charged peaks of a number of proteins to be distinctly separated. Application of MCR to a DESI image of a rat brain coronal section deconvoluted the image into components that showed biologically important features. Further application of MCR to a subsection of the image produced a component that clearly separated out the substantia nigra region, which allowed us to produce a biochemical anatomy for this area of the brain. We have demonstrated the ability of imaging MVA to automatically and objectively analyse DESI images of standardised and complex biological samples, and have shown its capacity for detailed spatial profiling of biomolecules in specific morphological regions. We propose the routine use of this technique for future DESI imaging experiments.