Non-targeted analysis (NTA), using liquid chromatography coupled with high resolution mass spectrometry (LC-HRMS), is an approach for the characterization of chemicals within complex samples, containing thousands of known and un-known constituents. LC-HRMS allows for simultaneous separation (i.e., reduction of sample complexity) and information collection on the chemical structure (i.e., mass spectra). One of the data processing steps taken after data collection and before chemical identification is componentization. This step groups all individually detected features into components in an attempt to represent single chemical constituents.  Currently available componentization algorithms have two major bottle necks related to the isotope detection and fragment detection. Isotope detection either requires information on the chemical formula of the feature, which is unknown during NTA, or a mass difference of 1.0033 Da is used for identification. The latter approach only covers part of all potential isotopes that can be found in mass spectra. As for the fragment ion detection, only chromatographic information (i.e., retention time and peak shape) is utilized. Due to the lack of knowledge on the chemical structure it becomes difficult to assess whether parent and fragment ions belong to the same chemical constituent. Finally, none of the current algorithms are able to combine information of both the MS1 (i.e., adducts, isotopes and in-source fragments) and MS2 (i.e., fragments) level.  To overcome these current limitations the componentization algorithm “CompCreate” was developed. Isotopes are identified using the Kendrick mass defect, allowing for the detection of all potential isotopes without the need for structural information. As for the fragment ions, a neutral loss database was constructed based on reference spectra (MassBank EU), allowing for the assessment of the occurrence likelihood of a neutral loss between the parent and fragment ions. Evaluation of the algorithm, using internal standards at different concentration, showed a false detection rate below 10%. Additionally, component identification for complex wastewater influent samples showed that 1.2% of the components were wrongly identified with a similar compound, while 7.3% was wrongly identified due to the lack of ions present. Overall, CompCreate showed a great potential for the componentization of LC-HRMS data (i.e., both MS1 and MS2).

Probabilistic approach for the componentization of non-targeted liquid chromatography high-resolution mass spectrometry data

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.