Quantification for non-targeted LC/MS screening without standard substances.

Area of Science:

• Analytical Chemistry
• Environmental Science
• Food Safety

Background:

• Non-targeted and suspect analyses using liquid chromatography/electrospray/high-resolution mass spectrometry (LC/ESI/HRMS) are crucial for identifying numerous compounds in complex samples.
• Quantifying these identified compounds accurately without authentic standards presents a significant analytical challenge.
• Existing methods often require costly and time-consuming standard preparation for each analyte.

Purpose of the Study:

• To develop and validate a novel computational approach for quantifying compounds identified via LC/ESI/HRMS, even in the absence of reference standards.
• To assess the accuracy and transferability of predicted compound responses across different analytical instruments.
• To enable reliable concentration estimations for environmental and food safety applications.

Main Methods:

• Utilized random forest regression to predict the electrospray ionization (ESI)/high-resolution mass spectrometry (HRMS) response of compounds.
• Developed a regression-based method for transferring predicted responses between different LC/ESI/HRMS instruments.
• Applied predicted responses to estimate compound concentrations in real-world samples without using authentic standards.

Main Results:

• Achieved a mean error of 2.2 (positive mode) and 2.0 (negative mode) times for predicted ESI/HRMS responses.
• Demonstrated successful transferability of predicted responses across different instruments.
• Validated the approach by quantifying pesticides and mycotoxins in cereal samples with an average quantification error of 5.4 times, compatible with toxicological predictions.

Conclusions:

• The developed random forest regression approach provides a robust method for quantifying compounds from LC/ESI/HRMS data without authentic standards.
• The predicted responses are instrument-transferable, enhancing the method's practical applicability.
• This technique offers a viable solution for accurate compound quantification in food safety and environmental monitoring, aligning with the accuracy needed for effect predictions.