Comparison of two algorithmic data processing strategies for metabolic fingerprinting by comprehensive two-dimensional gas chromatography–time-of-flight mass spectrometry

The alignment algorithm Statistical Compare (SC) developed by LECO Corporation for the processing of comprehensive two-dimensional gas chromatography–time-of-flight mass spectrometry (GC × GC–TOFMS) data was validated and compared to the in-house developed retention time correction and data alignment tool INCA (Integrative Normalization and Comparative Analysis) by a spike-in experiment and the comparative metabolic fingerprinting of a wild type versus a double mutant strain of Escherichia coli (E. coli). Starting with the same peak lists generated by LECO's ChromaTOF software, the accuracy of peak alignment and detection of 1.1- to 4-fold changes in metabolite concentration was assessed by spiking 20 standard compounds into an aqueous methanol extract of E. coli. To provide the same quality input signals for both alignment routines, the universal m/z 73 trace of the trimethylsilyl (TMS) group was used as a quantitative measure for all features. The performance of data processing and alignment was evaluated and illustrated by ROC curves. Statistical Compare performed marginally better at the lower fold changes, while INCA did so at the higher fold changes. Using SC, quantitative precision could be improved substantially by exploiting the signal intensities of metabolite-specific unique (U) m/z ion traces rather than the universal m/z 73 trace. A list of 56 features that distinguished the two E. coli strains was obtained by the SC alignment using m/z U with an estimated false discovery rate (FDR) of <0.05. Ultimately, 23 metabolites could be identified, one additional and five less than with INCA due to the failure of SC to extract unitized m/z U's across all fingerprints with suitable spectral intensities for the latter metabolites.