MetScribeR: A Semiautomated Tool for Data Processing of In-House LC-MS Metabolite Reference Libraries

One persistent challenge in untargeted metabolomics is the identification of compounds from their mass spectrometry (MS) signal, which is necessary for biological data interpretation. This process can be facilitated by building in-house libraries of metabolite standards containing retention time (RT) information, which is orthogonal and complementary to large, published MS/MS spectra repositories. Creating such libraries can require substantial effort and is time intensive. To streamline this process, we developed metScribeR, an R package with a Shiny application to accelerate the creation of RT and m/z libraries. metScribeR provides an easy, user-friendly interface for peak finding, filtering, and comprehensive quality review of the MS data. Uniquely, metScribeR does not require MS/MS spectral information and reports an identification probability estimate for each adduct. In our benchmarking, metScribeR required approximately 10 s of computational and manual effort per standard, showed a correlation of 0.99 between manual and metScribeR-derived RTs, and appropriately filtered out poor quality peaks. The metScribeR output is a.csv file including the identity, m/z, RT, and peak quality information for standards along with MS/MS spectra retrieved from MassBank of North America (MoNA). metScribeR is open source and available for download on GitHub at https://github.com/ncats/metScribeR.