Msmls library of standards

Run alignment, peak picking, adduct deconvolution and feature annotation were sequentially performed on Progenesis QI v2.3 (Nonlinear Dynamics, Waters, Newcastle upon Tyne, UK). Detected peaks were annotated with regard to a set of pure reference standards (MSMLS Library of Standards, Sigma-Aldrich) measured under the experimental conditions described previously (Pezzatti et al., 2019b (link)). The following tolerances were used: 2.5 ppm for precursor and fragment mass, 10% for retention time (Rt), and 5% in the case of collisional cross section (CCS). Data processing was achieved by SUPreMe, which is in-house software with capabilities for drift correction, noise filtering and sample normalization. Finally, data were transferred to SIMCA-P 15.0 software (Umetrics, Umea, Sweden) for multi-variate analysis (MVA).

Chromatogram alignment, peak picking, adduct deconvolution, and feature annotation were sequentially performed on Progenesis QI v2.3 (Nonlinear Dynamics, Waters, Newcastle upon Tyne, UK). The following tolerances were used for feature annotation with regard to a set of pure reference standards (MSMLS Library of Standards, Sigma-Aldrich) measured in the same instrument: 2.5 ppm for precursor and fragment mass, 10% for Rt, and 5% in the case of CCS. Data pretreatment was performed with SUPreMe, an in-house software with capabilities for drift correction, noise filtering, and sample normalization. Finally, data were transferred to SIMCA-P 15.0 software (Umetrics, Umea, Sweden) to perform Principal Component Analysis (PCA). AMOPLS analysis was conducted after unit variance scaling as previously described [42 (link)] under the MATLAB^® 8 environment (The MathWorks, Natick, MA, USA). A series of 10⁴ random permutations was performed to validate the AMOPLS model and assess the statistical significance of the effects.