From noise to signal: Kendrick mass filtering for high-resolution mass spectrometry imaging analysis

Introduction
Over the last years, lots of progress have been done in the development of mass spectrometry imaging, making the technique more and more accessible for various applications, such as biomarkers discovery or bioactive compounds identification. However, the progresses made in terms of spatial and instrumental resolution has for consequences the dramatic increase of dataset size, shifting the burden from data production to data analysis. We propose here to use a semi-targeted method based on Kendrick mass defect (KMD) analysis to immediately identify the chemistry-related compounds in mass spectrometry imaging applied to microbiology samples. Thanks to an in-house software, we are now able to better understand the bacteria-bacteria interactions.

Materials and methods
Bacillus velezensis GA1 and Pseudomonas sp. CMR12a were inoculated on a semi-solid agar-based medium and incubated at 30°C. Region of interest was cut directly from the petri dish and transferred to the target ITO plate, previously covered with double sided conductive carbon tape. This assembly was then put in a vacuum desiccator until complete drying (overnight), and covered with HCCA matrix. Mass spectrometry images were obtained using a FT-ICR mass spectrometer (9.4T SolariX, Bruker Daltonics, Bremen, Germany). Data analysis was performed on an in-house software.

Results & Discussion
Thanks to the KMD analysis, we were able to directly compare and identify the nature of the compounds detected in MSI such as lipids (1) or lipopeptides (2a), without the need of an extensive database search. It was also possible to identify some lipopeptides degradation occurring nearby Pseudomonas (2b). Thanks to our in-house software, the compounds with a similar chemistry can now be filtrated and the image can be reconstructed, removing thus the noise and focusing only on the signal.