HREI-MSDB: Data Science-Ready Database of High-Resolution Electron Ionization Mass Spectra

Abstract Despite the increasing use of gas chromatography coupled with electron ionization high-resolution mass spectrometry (GC-EI-HRMS), only a limited amount of GC-EI-HRMS reference data is currently available. Such data are beneficial not only for use with a GC-EI-HRMS instrument but also for computational purposes─for example, for training machine learning models that predict fragmentation under electron ionization conditions. Such models are currently trained and tested on integer m/z data, although they predict the exact elemental composition of ions. High-quality GC-EI-HRMS data will be able to boost the development of such models even if the end user does not have such an instrument. We present HREI-MSDB: a GC-EI-HRMS mass spectral library with an m/z range of 40–500, and a resolution (full width at half-maximum) of 30,000. The error in m/z determination for almost all peaks (except for special cases such as overlapping of intense peaks) does not exceed 0.6 mDa; typical values are 0.1–0.4 mDa or 0.5–1.5 ppm. The data set includes 329 mass spectra of molecules of various classes (nitrogen-containing heterocycles, aromatic and polyaromatic compounds, alcohols, phenols, and essential oil components). All mass spectra are manually verified, free of intense impurity peaks, and are of high quality. In the future, mass spectra may be posted on MassBank. The data is available online: https://doi.org/10.6084/m9.figshare.29713460