Comparing mass, retention time and tandem mass spectra as criteria for the automated screening of small molecules in aqueous environmental samples analyzed by liquid chromatography/quadrupole time-of-flight tandem mass spectrometry.

RATIONALE: The adoption of database screening using high-resolution liquid chromatography/mass spectrometry data is promising as a river water monitoring and surveillance tool but depends on the ability to perform reliable data processing on a large number of samples in a unified workflow. Strategies to minimize errors have been proposed but automated procedures are rare. METHODS: High-resolution LC/ESI-QTOFMS/MS in data-dependent MS2 acquisition mode was performed for the analysis of surface water samples by direct injection. Data processing was achieved with software tools written in R. A database containing MS2 spectra of 693 compounds formed the basis of the workflow. Standard mixes and a time series of 361 samples of river water were analyzed and processed with the optimized workflow. RESULTS: Using the database and a mix of 70 standards for testing, it was found that an identification strategy including (i) mass, (ii) retention time, and (iii) MS2 spectral matching achieved a two- to three-fold improvement in the fraction of false positives compared with using only two criteria, while the number of false negatives remained low. The optimized workflow was applied to the sample series of river water. In total, 135 compounds were identified by a library match. CONCLUSIONS: The developed automated database screening approach minimizes the proportion of false positives, while still allowing for the screening of hundreds of water samples for hundreds of compounds in a single run.