Figure 1.

Workflow of CP-Hunter. Profile-mode MS data are translated into centroids (1). Theoretical isotopic distributions of CPs and their transformation products are obtained from ChemCalc library (2).²⁸ Interfering signals are deconvoluted from one another by non-negative matrix factorization (3).²⁹ Carbon- and chlorine-homologue abundances (MI_100%-values) are calculated based on all signals (I_100%-values) of the isotope cluster (4). Evaluation of deconvoluted isotope clusters is recommended for low-abundant compounds to identify false positive signals (5). Reconstructed mass spectra (RMS) containing assigned signals to CPs, COs, CdiOs and CtriOs compound classes and not assigned signals are obtained (6).