Quantitative Evaluation of Algorithms for Isotopic Envelope Extraction via Extracted Ion Chromatogram Clustering

LC–MS
precursor (MS1) data are used increasingly often in
conjunction with MS/MS data for the quantification, validation, and
other computational mass spectrometry tasks. The efficacy of MS1 data
on downstream tasks is dependent on the coverage and accuracy of the
MS1 isotopic envelope extraction algorithms that delineate them from
the dense backgrounds common in complex samples. Although several
algorithms for extracted ion chromatogram (XIC) clustering exist,
their performance has not yet been quantified, in part due to the
difficulty of obtaining, isolating, and running some algorithms and
in part due to the lack of quantitative MS1 ground truth. Using a
newly available manually annotated ground truth data set, we measure
the performance of several popular XIC clustering algorithms in time,
coverage, and accuracy of resulting isotopic envelopes. We intend
this work to provide a benchmark against which future algorithms can
be scored.