Source-linked AI summary
MALDIquant: a versatile R package for the analysis of mass spectrometry data
Sebastian Gibb, Korbinian Strimmer
TL;DR
Clinical mass spectrometry profiling requires sophisticated statistical analysis, particularly when spectra must be compared quantitatively. MALDIquant provides an extensible R-based workflow covering import, preprocessing, alignment, calibration, and downstream analysis, with support for technical replicates and unequal-resolution spectra. The package's central capabilities are non-linear peak alignment and cross-spectrum peak-intensity calibration.
Problem
Clinical mass spectrometry profiling generates high-throughput data requiring sophisticated statistical algorithms for analysis.
Method
MALDIquant implements a complete, extensible R analysis pipeline for importing, preprocessing, detecting and aligning peaks, calibrating spectra, and exporting matrices for statistical analysis.
Results
MALDIquant provides non-linear peak alignment, peak-intensity calibration across spectra, and analysis of technical replicates and unequal-resolution spectra.
Takeaways & Limitations
MALDIquant offers a flexible pipeline for MALDI-TOF and other mass spectrometry data, including clinical-proteomics workflows and downstream classification or feature selection.
Abstract
from arXiv · showhide
Summary: MALDIquant is an R package providing a complete and modular analysis pipeline for quantitative analysis of mass spectrometry data. MALDIquant is specifically designed with application in clinical diagnostics in mind and implements sophisticated routines for importing raw data, preprocessing, non-linear peak alignment, and calibration. It also handles technical replicates as well as spectra with unequal resolution. Availability: MALDIquant and its associated R packages readBrukerFlexData and readMzXmlData are freely available from the R archive CRAN (http://cran.r-project.org). The software is distributed under the GNU General Public License (version 3 or later) and is accompanied by example files and data. Additional documentation is available from http://strimmerlab.org/software/maldiquant/.
1 Introduction
Mass spectrometry profiling is increasingly important in clinical diagnostics, but its high-throughput data require sophisticated statistical analysis. MALDIquant addresses this need with a complete, extensible R-based pipeline initially developed for clinical proteomics but applicable to other two-dimensional mass spectrometry data.
- Mass spectrometry profiling is increasingly used in clinical diagnostics, including biomarker identification for cancer.
- MALDIquant provides an open-source R pipeline spanning raw-data import, preprocessing, peak detection, non-linear alignment, and mass-spectra calibration.Its standalone S4 object-oriented design supports further extension.
- Although initially developed for MALDI-based clinical proteomics, MALDIquant's algorithms may also be applied to other 2D mass spectrometry data.
2 Distinctive Features
MALDIquant distinguishes itself from related R packages through non-linear peak alignment, cross-spectrum intensity calibration, and support for technical replicates and unequal-resolution spectra.
- MALDIquant implements sophisticated non-linear peak alignment and calibration of peak intensities across spectra.The calibration procedure is modeled on a method for sequence count data.
- The package supports analysis of technical replicates and spectra with unequal resolution, enabling comparisons of spectra from multiple clinical sources.
3 Details on Algorithms
MALDIquant's workflow imports and preprocesses spectra, aligns peaks through non-linear warping, calibrates intensities, and exports matrices for downstream statistical analysis. Its routines include smoothing, baseline correction, peak detection, replicate merging, and classification-oriented feature selection.
- 3 Details on Algorithms: The example workflow progresses from raw spectra through smoothing, baseline correction, peak detection, warping-based alignment, and merged-spectrum visualization.Figure 1 labels these stages A through F.
- 3.1 Data import: MALDIquant supports hardware-independent analysis while providing associated packages for Bruker *flex binary files, folder hierarchies, and mzXML input.
- 3.2 Data preprocessing: Preprocessing includes smoothing, variance stabilization, baseline correction, and peak detection; the default SNIP baseline method produces positive corrected intensities.
- 3.3 Peak alignment: Peak alignment identifies landmark peaks and fits a non-linear warping function by local regression to matched reference peaks.The procedure also permits merging aligned technical-replicate spectra.
- 3.4 Calibration: Calibration creates a reference spectrum from median aligned-peak intensities and computes each spectrum's scale factor using a robust overall intensity-ratio estimator.Calibration based on total ion current is also available.
- 3.5 Classification and feature selection: The calibrated peak-intensity matrix can be exported for classification and feature selection using shrinkage discriminant analysis.
4 Conclusion
MALDIquant is presented as a versatile and flexible R package for MALDI-TOF and other mass spectrometry data. Its demonstration script provides an overview of its capabilities.
- MALDIquant provides a flexible analysis pipeline for MALDI-TOF and other mass spectrometry data, with non-linear warping alignment and simultaneous peak-intensity calibration.
- The included demo script can be run to overview MALDIquant's capabilities.The documented commands load the package and invoke its demo.