Nontargeted Screening of Water Samples Using Data-Dependent Acquisition with Similar Partition Searching

A rapid screening method for the detection of nontargeted compounds in surface water samples was developed using MS-MS high-resolution mass spectrometry and data-dependent acquisition. The key parameters for the acquisition method were optimized using five model compounds with diverse chemical characteristics. The parameter selection required optimization between the total number of precursor ions that could be selected in an LC-MS run, the quality of each MS (full range) spectrum, and the quality of each MS-MS fragmentation spectrum. After the acquisition method was optimized, 18 surface water samples from rivers, reservoirs, and effluents from wastewater treatment plants were analyzed, generating 41625 MS-MS spectra in about 14 h. The raw data were then converted into two generic formats using the open-access program MSConvert. A combinatorial approach, similar partition searching (SPS), was then used to putatively identify analytes from the accurate mass of each analyte (adjusted for the adduct mass) and the corresponding MS-MS spectra were obtained. In this approach, the structures of about 250000 common compounds, stored in a large database as mathematical partitions of their exact mass, were compared directly to each MS-MS spectrum. Compounds with a similar mass and retention time were grouped together and labeled as "Analytes", using an Excel Add-In. The isotope ratio data from the MS spectrum, the corresponding MS-MS spectra, and the putative identifications were then imported into an Access relational database to facilitate sorting, searching, filtering, and querying the results. This allowed final inspection to assess confidence of the identifications made through the nontargeted screening.