MSACL 2017 US Abstract

Data Dependent or Data Independent Acquisition: Evaluation of SWATH for Drug Screening

Kara Lynch (Presenter)
University of California San Francisco

Authorship: Kara L. Lynch
University of California San Francisco

Short Abstract High resolution mass spectrometry has become increasingly common for drug screening in clinical and forensic toxicology laboratories. Data-dependent strategies are commonly used for automated data acquisition (DDA). For this study two data-independent acquisition (DIA) methods using Sequential Windowed Acquisition of all Theoretical fragment-ion spectra (SWATH) were developed (fixed and variable window isolation) and compared to a DDA method using a QTOF mass spectrometer. Limits of detection, matrix effects, and the ability to identify 115 drugs/metabolites in 50 patient samples, using the three methods, were evaluated. SWATH was deemed comparable to DDA with increased detection for some low abundance compounds.

Long Abstract

Introductions: In recent years, high resolution mass spectrometry has become increasingly common for drug screening in clinical and forensic toxicology laboratories. Data-dependent strategies are commonly used for automated data acquisition (DDA). For this study two data-independent acquisition (DIA) methods using Sequential Windowed Acquisition of all Theoretical fragment-ion spectra (SWATH) were developed (fixed and variable window isolation) and compared to a DDA method using a QTOF mass spectrometer. SWATH performs data-independent fragmentation of all precursor ions entering the mass spectrometer in specified isolation windows covering the specified mass range. This allows multiple repeat analyses of each window during the elution of a single chromatographic peak. With DDA, the Q1 mass isolation window for collection of product ion spectra is typically less than 1 Da, however with DIA the window is larger (~20 Da) Product ions are collected for all ions within the window resulting in a “less-pure” spectra compared to DDA spectra, however, data is collected for all ions. Low abundance ions can be missed using DDA due to the limited number of triggered product ion scans at any given time (typically 10-20). The objective of this study was to compare DDA and DIA methods for the detection of drugs and metabolites in biological samples.

Methods: For all methods, urine samples were diluted 1:10 and separation was performed using a Phenomenex Kinetex C18 column (50x3.00mm, 2.6μm) with a 10 minute gradient from 2%-100% organic (MPA – 5mM ammonium formate, 0.05% formic acid and MPB – 50:50 methanol:acetonitrile, 0.05% formic acid). Data was collected on a SCIEX TripleTOF®5600 operating in positive-ion mode with Analyst TF software using: 1) TOF-MS survey scan from 50-700 Da with DDA-triggered collection of up to 20 product ion scans at a time (DDA method), 2) SWATH acquisition from 100-650 Da using 30 fixed 18 Da windows (SWATH-fixed method), and 3) SWATH acquisition from 100-650 Da using 30 variable windows ranging from 6 – 59 Da (SWATH-variable method). Data analysis was done using PeakView® software (SCIEX) and an in-house generated mass spectral library. A combined score for positive identification was computed based on 4 confidence categories (mass error, retention time error, isotope ratio difference, and library score) with custom weighting determined experimentally for each method. LOD was determined by spiking drug standards into drug-free urine at various concentrations (5, 10, 25, 50, 100, 125, 250, and 500 ng/mL). LOD for 115 drugs and metabolites was defined as the lowest concentration for which the drug was called positive by the scoring criteria and the signal-to-noise ratio was >20:1. Matrix effects, using the three methods, were determined by spiking 115 drug standards into 3 matrices at two concentrations and injecting in triplicate. Matrix effects were calculated using the following equation: (B – A)/A X 100%, where B is the mean signal intensity in urine and A is the signal intensity in water. Remnant well characterized urine samples (N = 50) were tested using the three methods. All information about these samples (immunoassay, GC-MS, LC-MS/MS, LC-Qrbitrap results and patient prescription history) was used to confirm discordant results.

Results: The limit of detection for all 115 drugs and metabolites evaluated was comparable between the three methods using the spiked LOD samples, with >70% of compounds detected at less than 25 ng/mL. The SWATH methods were slightly more sensitive, but in many cases, this was a difference of only 5 ng/mL or 15 ng/mL. Matrix effects were observed and similar for the three methods, which is to be expected given the same liquid chromatograpy and ion source conditions were used for the three methods. The detection capabilities in the 50 remnant samples were similar for the three methods, however, the SWATH-variable method detected a few additional low abundance drugs and metabolites (N-28) compared to the DDA method. For, these compounds, a peak was detected using DDA, however, no product ion was acquired and thus reporting criteria were not met. In many of these cases another metabolite of the drug was identified using the DDA method therefore overall identification was not significantly different between the method.

Conclusions: SWATH is a viable alternative to DDA methods and in many cases resulted in more sensitive detection of low abundance ions; however, the data review was more time-consuming.

References & Acknowledgements:

Financial Disclosure

IP Royalty: no

Planning to mention or discuss specific products or technology of the company(ies) listed above:

no