Peak Detection Method Evaluation for Ion Mobility Spectrometry by Using Machine Learning Approaches

Abstract

Ion mobility spectrometry with pre-separation by multi-capillary columns (MCC/IMS) has become an established inexpensive, non-invasive bioanalytics technology for detecting volatile organic compounds (VOCs) with various metabolomics applications in medical research. To pave the way for this technology towards daily usage in medical practice, different steps still have to be taken. With respect to modern biomarker research, one of the most important tasks is the automatic classification of patient-specific data sets into different groups, healthy or not, for instance. Although sophisticated machine learning methods exist, an inevitable preprocessing step is reliable and robust peak detection without manual intervention. In this work we evaluate four state-of-the-art approaches for automated IMS-based peak detection: local maxima search, watershed transformation with IPHEx, region-merging with VisualNow, and peak model estimation (PME). We manually generated a gold standard with the aid of a domain expert (manual) and compare the performance of the four peak calling methods with respect to two distinct criteria. We first utilize established machine learning methods and systematically study their classification performance based on the four peak detectors’ results. Second, we investigate the classification variance and robustness regarding perturbation and overfitting. Our main finding is that the power of the classification accuracy is almost equally good for all methods, the manually created gold standard as well as the four automatic peak finding methods. In addition, we note that all tools, manual and automatic, are similarly robust against perturbations. However, the classification performance is more robust against overfitting when using the PME as peak calling preprocessor. In summary, we conclude that all methods, though small differences exist, are largely reliable and enable a wide spectrum of real-world biomedical applications.

From raw ion mobility measurements to disease classification: a comparison of analysis processes

“…Grid Square This fast and simple approach partitions the coordinate system of consensus peaks into disjoint rectangular regions. Each region corresponds to a potential consensus peak and has fixed width and linearly increasing height depending on the retention time, as stated by Hauschild et al (2013). For every peak, its associated region can be computed in constant time.…”

“…Runs with more than one change of the sign of the derivative are discarded. Runs are merged into peaks over different retention time when certain distance criteria stated by Hauschild et al (2013) are fulfilled.…”

“…Additionally, the effect of interactions between different algorithms for all three steps of the process on the final classification results has not been studied in detail. The work by Hauschild et al (2013) does study a similar question, but naturally does not cover recently developed peak detection methods and also only a smaller selection of classifiers.…”

Application of multi-capillary column – ion mobility spectrometry (MCC-IMS) in rubber chemistry and technology

“…1 Scheme of ion mobility spectrometer [3] To make IMS more effective for analysis of volatile organic compounds, the IMS instrumentation is often coupled with a multi capillary (MCC) or gas chromatography column (GC). Sample is pre-separated on MCC and then an analyte goes to IMS directly.…”

An Online Peak Extraction Algorithm for Ion Mobility Spectrometry Data

“…The approach writes F =s·G +(1−s)·H withs ∈ [ 0, 1], where H is a non-parametric distribution whose inclusion ensures the fit of the model G to the data F. If the weights is close to 1.0, then F is a plausible sample from G. We compare the original spectra and reduced spectra (peaks from peak lists) from a previously used dataset [20]. This set contains 69 measurements preprocessed with a 5 × 5 average.…”