Non-target screening: quantitative and on time

However, to date identification of compounds from NTS data is challenging due to the vast chemical space covered by the potential candidate contaminants. Moreover, NTS is not inherently quantitative; compound intensities cannot directly be translated into compound concentrations due to differing ionization efficiencies. Consequently, quantification of OMPs currently relies on the use of standards. The overall aim of this project is to address these two issues and thereby reduce the gap between target analyses and non-target screening; we therefore propose to 1) improve compound identification in LC-HRMS based NTS through candidate filtering based on retention time indices, and 2) to enable the quantification of suspects without using standards.

Improvement and extension of NTS workflows for water contaminant identification

Liquid chromatography high resolution mass spectrometry (LC-ESI-HR-MS) can combine target and non-target analyses, given chromatographic peak identification is confident and quantitative. The proposed project will work towards that goal through improvement and extension of existing NTS workflows (Figure 1).

Schematics of non-target screening based identification and quantification

More confident identification based on retention time indices and quantification without the need of standards

The implementation of candidate substance (suspects) filtering based on chromatographic retention time (Activity 1) can result in a significant reduction in potential suspect numbers, lower false positive rates and more confident feature identification. To this end, algorithms for retention time indices (RTI) will be compared and their performance tested experimentally to arrive at an optimal index. The selected index will then be implemented into existing NTS workflows. Furthermore, the use of an RTI can facilitate NTS data exchange between laboratories.

As there is a generic target value with an alerting function set at 1 µg/L for “other” organic anthropogenic chemicals in sources of drinking water in The Netherlands, the intensity of an NTS feature needs to be translated into a concentration. In this project, suspect concentrations will be quantified without the need for standards based on ionization efficiencies with the Kruve method (Activity 2). Kruve et al investigated how the structure of a compound, the eluent used in the analysis, the sample matrix and the instrument used for the analysis influence ionization efficiency. (Liigand et al., 2014; Liigand et al., 2017; Ojakivi et al., 2018). In general, the more hydrophobic compounds and more basic compounds tend to have a higher response in ESI positive mode. Higher organic solvent content and lower pH result in higher ionization efficiency and, therefore, a higher response in positive mode ESI/MS. Moreover, the ionization efficiency values measured on one instrument are transferable to other instruments with the aid of 5-6 common compounds (Liigand et al., 2015). Combining these results and machine learning approaches Kruve et al. developed an universal approach for applying ionization efficiency predictions for NTS feature quantification Figure 2). We will collaborate with Dr. Kruve and extend their method to water matrix samples and to C18, HILIC and Mixed mode chromatographies.

Improved chemical water quality monitoring is key for risk based monitoring strategies

The proposed project addresses current shortcomings in the NTS and therefore makes an important contribution to monitoring chemical water quality and risk-based monitoring strategies. In addition, this project supplements the Dutch Technical Agreement (NTA) “Guideline for Non Target Screening of organic substances in water samples with chromatography and mass spectrometry” that is currently being drawn up together with drinking water laboratories and NEN.

As part of the dissemination, manuscripts on the results of Activity 1 and 2 will be prepared for scientific publication. To facilitate implementation at the drinking water laboratories, a workshop will be held on the Kruve method and RTI indices and the developed tools added to existing NTS workflows where possible.