Power of mzRAPP-Based Performance Assessments in MS1-Based Nontargeted Feature Detection

When performing chromatography-mass spectrometry-based nontargeted metabolomics, or exposomics, one of the key steps in the analysis is to obtain MS1-based feature tables. Inapt parameter settings in feature detection will result in missing or wrong quantitative values and might ultimately lead to downstream incorrect biological interpretations. However, until recently, no strategies to assess the completeness and abundance accuracy of feature tables were available. Here, we show that mzRAPP enables the generation of benchmark peak lists by using an internal set of known molecules in the analyzed data set. Using the benchmark, the completeness and abundance accuracy of feature tables can be assessed in an automated pipeline. We demonstrate that our approach adds to other commonly applied quality assurance methods such as manual or automatized parameter optimization techniques or removal of false-positive signals. Moreover, we show that as few as 10 benchmark molecules can already allow for representative performance metrics to further improve quantitative biological understanding.

Automated summary

When performing chromatography-mass spectrometry-based nontargeted metabolomics or exposomics, evaluating the completeness and accuracy of feature tables is crucial. The newly developed mzRAPP method allows the generation of benchmark peak lists using known molecules, facilitating the improvement of quantitative biological understanding.