Refine
Keywords
Year of publication
- 2020 (1)
Document Type
- Article (1)
Language
- English (1)
Has Fulltext
- no (1)
MALDI mass spectrometry imaging (MALDI MSI) is a spatially resolved analytical tool for biological tissue
analysis by measuring mass-to-charge ratios of ionized molecules. With increasing spatial and mass resolution of
MALDI MSI data, appropriate data analysis and interpretation is getting more and more challenging. A reliable
separation of important peaks from noise (aka peak detection) is a prerequisite for many subsequent processing
steps and should be as accurate as possible. We propose a novel peak detection algorithm based on sparse frame
multipliers, which can be applied to raw MALDI MSI data without prior preprocessing. The accuracy is evaluated
on a simulated data set in comparison with state-of-the-art algorithms. These results also show the proposed
method's robustness to baseline and noise effects. In addition, the method is evaluated on real MALDI-TOF data
sets, whereby spatial information can be included in the peak picking process.
Significance: The field of proteomics, in particular MALDI Imaging, encompasses huge amounts of data. The
processing and preprocessing of this data in order to segment or classify spatial structures of certain peptides or
isotope patterns can hence be cumbersome and includes several independent processing steps. In this work, we
propose a simple peak-picking algorithm to quickly analyze large raw MALDI Imaging data sets, which has a
better sensitivity than current state-of-the-art algorithms. Further, it is possible to get an overall overview of the
entire data set showing the most significant and spatially localized peptide structures and, hence, contributes all
data driven evaluation of MALDI Imaging data.