Improving annotation of known-unknowns with accurately reconstructed mass spectra

Chemical profiling with gas chromatography/mass spectrometry (GC/MS) full-spectrum acquisition mode often leads to the discovery of known-unknown components. These are non-identified components which arise from the limitation of the data processing method or limited breadth of mass spectral libraries. The recent introduction of the NIST Hybrid Search sought to relieve the latter limitation by providing an improved class annotation technique. Herein, we demonstrate the importance of using a precise mass spectral reconstruction technique to increase the confidence of detecting the presence of known-unknowns and subsequently identifying them successfully with NIST Hybrid Search function. We compared the AMDIS algorithm against the rBTEM algorithm and found that the latter could more accurately reconstruct the mass spectra of co-eluting known-unknown components. This has a far-reaching implication to increase the number of identified compounds in GC/MS scan data.