Fragmentation Trees for Automated de novo Interpretation of Impure Electron Ionization Spectra from Gas Chromatographic Complex Mixture Analysis—Chemical Deconvolution

-Impure spectra contain signals from multiple analytes or signals unrelated to the main analyte, due to propinquitous coelution that confounds purely mathematical de-coelution. Some such spectra are expected in complex sample analysis.
-Impure spectra may impede manual or automated interpretation and beget false assignments or other failures.
-Chemical information within high-resolution spectra can assist mathematical de-coelution and facilitate analyte identification.