Extended Data Fig. 3: ZODIAC processing and evaluation workflow. | Nature Machine Intelligence

Extended Data Fig. 3: ZODIAC processing and evaluation workflow.

From: Database-independent molecular formula annotation using Gibbs sampling through ZODIAC

Extended Data Fig. 3

(1) Each LC-MS/MS run is processed individually; input mzML/mzXML files are processed using OpenMS, performing feature and adduct detection and producing files in SIRIUS input format. Resulting features combine MS1, MS/MS and adduct information. (2), (3) Filtering is performed on feature, MS/MS and peak level. (4) Similar features are merged between different runs using hierarchical clustering; MS/MS are combined and a best isotope pattern is selected per feature. (5) Missing isotope peaks are searched in MS1 spectra to extend isotope patterns. (6) A final feature filtering step is performed; the remaining features are considered as compounds. (7) SIRIUS is executed. (8) Compounds with few explained peaks are discarded, since a badly explained MS/MS spectrum indicates low quality. (9) ZODIAC is run on the remaining compounds. (10) SIRIUS and ZODIAC are evaluated on the same set of compounds.

Back to article page