Targeted vs untargeted MS2 data-dependent acquisition for automated peak annotation in LC-MS metabolomics

The data set includes MS and MSMS data collected from the analysis of human milk samples, using four different strategies for automated MS2 DDA, which can be easily implemented in the frame of standard QA/QC workflows for untargeted liquid chromatography-mass spectrometry (LC-MS). These strategies include (i) DDA in the MS working range; (ii) iterated DDA split into several m/z intervals; (iii) dynamic iterated DDA of (pre)selected potentially informative features; and (iv) dynamic iterated DDA of (pre)annotated metabolic features using a reference database (HMDB).
MS data sets: 2 blanks and a set of QCs were injected at the beginning of the sequence for system conditioning and MS2 data acquisition (data set: I0). Then, the sample batch including 42 milk samples, 13 QCs (1 QC every 6 samples) and 3 blanks were analyzed (data set: B0). Peak table generation was carried out using XCMS software. The centWave method was used for peak detection with the following parameters: mass accuracy, 20 ppm; peak width, (5,25); snthresh, 12; prefilter, (5,5000); minimum difference for overlapping peaks: 7.5 mDa; intensity weighted m/z values of each feature were calculated using the wMean function; Peak limits used for integration: Mexican hat filtered data. Grouping before and after RT correction was carried out using the nearest method and 9 s as rtCheck argument. Missing data points were filled by reintegrating the raw data files in the regions of the missing peaks using the fillPeaks method. The CAMERA package was used for the identification of pseudospectra based on peak shape analysis, isotopic information and intensity correlation across samples.
MSMS data sets: Two untargeted and two targeted DDA strategies for automated MS2 data acquisition based on the algorithm depicted in Figure 1B were employed: (i) untargeted selection of precursors in the 70-1500 Da range (DDA); (ii) untargeted iterated DDA, in which MS2 spectra were acquired in consecutive QC replicates using untargeted DDA in the [70-200], [200-400], [400-600], [600-800], [800-1000], [1000-1250], and [1250-1500] Da ranges (i-DDA); (iii) targeted dynamic iterated DDA, in which MS2 spectra were acquired by automated selection of precursor ions using an inclusion list generated after the injection of two blanks and three QCs during system conditioning (xcms-DDA). Raw MSMS data (.D) was converted into .ms2 and .mgf format using ProteoWizard. ms2 data was directly imported into MATLAB (ms2_dda, ms2_hmdb, ms2_idda and ms2_xcms data structures).