Theodore Alexandrov1, Andreas Bartels. 1. Center for Industrial Mathematics, University of Bremen, 28359 Bremen, Germany. theodore@uni-bremen.de
Abstract
MOTIVATION: Imaging mass spectrometry has emerged in the past decade as a label-free, spatially resolved and multi-purpose bioanalytical technique for direct analysis of biological samples. However, solving two everyday data analysis problems still requires expert judgment: (i) the detection of unknown molecules and (ii) the testing for presence of known molecules. RESULTS: We developed a measure of spatial chaos of a molecular image corresponding to a mass-to-charge value, which is a proxy for the molecular presence, and developed methods solving considered problems. The statistical evaluation was performed on a dataset from a rat brain section with test sets of molecular images selected by an expert. The measure of spatial chaos has shown high agreement with expert judges. The method for detection of unknown molecules allowed us to find structured molecular images corresponding to spectral peaks of any low intensity. The test for presence applied to a list of endogenous peptides ranked them according to the proposed measure of their presence in the sample. AVAILABILITY: The source code and test sets of mass-to-charge images are available at http://www.math.uni-bremen.de/∼theodore. SUPPLEMENTARY INFORMATION: Supplementary materials are available at Bioinformatics online. CONTACT: theodore@uni-bremen.de.
MOTIVATION: Imaging mass spectrometry has emerged in the past decade as a label-free, spatially resolved and multi-purpose bioanalytical technique for direct analysis of biological samples. However, solving two everyday data analysis problems still requires expert judgment: (i) the detection of unknown molecules and (ii) the testing for presence of known molecules. RESULTS: We developed a measure of spatial chaos of a molecular image corresponding to a mass-to-charge value, which is a proxy for the molecular presence, and developed methods solving considered problems. The statistical evaluation was performed on a dataset from a rat brain section with test sets of molecular images selected by an expert. The measure of spatial chaos has shown high agreement with expert judges. The method for detection of unknown molecules allowed us to find structured molecular images corresponding to spectral peaks of any low intensity. The test for presence applied to a list of endogenous peptides ranked them according to the proposed measure of their presence in the sample. AVAILABILITY: The source code and test sets of mass-to-charge images are available at http://www.math.uni-bremen.de/∼theodore. SUPPLEMENTARY INFORMATION: Supplementary materials are available at Bioinformatics online. CONTACT: theodore@uni-bremen.de.
Authors: Gaël Picard de Muller; Rima Ait-Belkacem; David Bonnel; Rémi Longuespée; Jonathan Stauber Journal: J Am Soc Mass Spectrom Date: 2017-09-14 Impact factor: 3.109
Authors: Kirill Veselkov; Jonathan Sleeman; Emmanuelle Claude; Johannes P C Vissers; Dieter Galea; Anna Mroz; Ivan Laponogov; Mark Towers; Robert Tonge; Reza Mirnezami; Zoltan Takats; Jeremy K Nicholson; James I Langridge Journal: Sci Rep Date: 2018-03-06 Impact factor: 4.379
Authors: Paolo Inglese; Gonçalo Correia; Zoltan Takats; Jeremy K Nicholson; Robert C Glen Journal: Bioinformatics Date: 2019-01-01 Impact factor: 6.937
Authors: Kyle D Bemis; April Harry; Livia S Eberlin; Christina R Ferreira; Stephanie M van de Ven; Parag Mallick; Mark Stolowitz; Olga Vitek Journal: Mol Cell Proteomics Date: 2016-01-21 Impact factor: 5.911