High-Confidence Characterization Of Small Molecule 'Unknowns': How Software Is Enabling Intelligent, High-Resolution Accurate Mass LC-MS/MS Data Acquisition

Contributed Commentary by Scott M. Peterman

June 17, 2019 | Liquid chromatography-tandem mass spectrometry (LC-MS/MS) has proven to be a powerful tool for small molecule analysis, providing highly accurate and precise characterization data on a broad range of analytes. However, the identification of so-called "unknown-unknowns"—unexpected but important sample components for which there are no spectral library data—is extremely challenging.

Many LC-MS/MS small molecule analysis workflows rely upon data dependent acquisition (DDA), which uses data from an initial full-scan MS stage to select precursor ions to fragment in a second stage of MS (MS²). This can greatly reduce the complexity of data processing as only one precursor ion is fragmented in each MS² scan. Despite this, the nature of DDA means that, even with the most exhaustive methods, there is no guarantee that all features of interest attributed to unknown-unknown sample components will be analyzed in the DDA MS² step.

DDA involves selecting precursor ions in decreasing order of intensity so lower intensity ions may be missed. To increase precursor feature sampling by augmenting DDA, a dynamic exclusion (DE) routine is used that places all precursor m/z values onto a list for a user-defined time (generally 60-70% of the average chromatographic peak width) to eliminate repetitive MS² acquisition. This increases the breadth and depth of precursor sampling within the chromatographic elution profile of endogenous compounds.   

Despite the DDA/DE acquisition strategy, the background matrix dramatically increases the number of features measured in any full scan mass spectra. Determining which features should be attributed to the background matrix versus compounds originating from the sample can be challenging. To obtain sufficient data to characterize unknown compounds, MS data acquisition methods often require exhaustive sampling techniques using multiple replicate injections. This can generate very large volumes of data, making analysis time-consuming and increasing the likelihood of false positives.

A major issue facing MS data acquisition is therefore how best to trigger MS² spectral acquisition only for ions of interest, avoiding ions produced from background compounds and those sampled in previous injections. One approach is to exclude background and matrix features by analyzing a blank sample and manually creating an exclusion list, although this is very labor-intensive. Similarly, processing raw data files to determine which precursors have been sampled in previous injections is also inherently inefficient, potentially requiring weeks to months of valuable analyst time. As such, there is a real need for automated strategies that improve profiling efficiency by ensuring the data available for target identification is of high quality and relevant only to the compounds of interest.

Novel software solutions improving profiling efficiency

Recently, software algorithms have been developed that improve the identification of small molecule unknowns by using automated inclusion and exclusion lists, guiding precursor selection and the collection of exhaustive, high-quality data on all compounds of interest. As a result, lower intensity analytes are not missed, while "unimportant" or background features are excluded.

LC-MS/MS exclusion lists are created by running a background or matrix sample to identify features that should be removed from higher orders of MS (MSⁿ) analysis. Precursor mapping can be used to select features that shall not trigger MS² when test samples are run. To further increase profiling efficiency, inclusion lists are created by performing precursor mapping on a representative test sample. Features common to both lists are subsequently excluded, unless the measured ion intensity differs by a user-defined ratio. This approach enables the collection of more meaningful data in fewer replicate injections and improves the characterization of low-abundance analytes.

Using inclusion and exclusion lists together requires high performance chromatography and high- resolution MS technology, as the strategy relies on comparing mass-to-charge (m/z) values at specific retention times. Reproducible chromatography and high-resolution MS are therefore essential to define precursor features when precursors may be isobaric or have minute differences in mass. Given the complexity of these workflows, optimal results are achieved by tailoring the acquisition routine to the dedicated small molecule profiling mass spectrometer, as data acquisition routines can be designed around specific system capabilities.

Some of the latest systems allow exclusions and inclusions lists to be updated in real time, while iterative re-injection schemes can be used to determine which ions to fragment in the next analysis. Compounds sampled in MSⁿ can be moved to the exclusion list for subsequent injections, and data acquisition is automatically stopped after exhaustive sampling.

Using this approach to increase sampling efficiency also improves data quality. By increasing the efficiency of precursor selection, these workflows enable the collection of higher resolution data and facilitate the collection of MS³ and MS⁴ spectra for compounds of interest. This additional information is essential for the structural elucidation of "unknown-unknowns" that require substructure matching. Collecting higher order tandem MS data enables a more accurate definition of a compound’s unimolecular decomposition pathways, giving greater confidence in the identification of unknowns.

Following the completion of his PhD at Texas A&M University in 2001, Dr. Scott M. Peterman joined Thermo Fisher Scientific and has held several positions from application chemist, program manager, to now Senior Global Marketing Manager focusing on the Thermo Scientific Orbitrap Tribrid mass spectrometer platform and associated additions. He can be reached at scott.peterman@thermofisher.com