Measurable Feature Prediction for Estimating Chemical Space Coverage in LC-ESI-HRMS Nontargeted Analysis.

Area of Science:

• Environmental chemistry
• Analytical chemistry
• Computational chemistry

Background:

• Chemical space is vast and expanding, necessitating methods to determine experimentally measurable regions.
• Nontargeted analysis (NTA) for exposomics and environmental monitoring requires understanding method-specific constraints like retention and ionization.
• Current NTA methods using LC-ESI-HRMS are limited in the chemical space they can access.

Purpose of the Study:

• To develop an in silico framework for predicting measurable chemical features and coverage under specific LC-ESI-HRMS conditions.
• To enable pre-analysis assessment of chemical space accessibility for NTA.
• To guide the selection of analytical methods for comprehensive environmental monitoring.

Main Methods:

• Integrated experimental data (internal standards) with molecular fingerprints and quantitative structure-property relationships (QSPR).
• Utilized large-scale chemical databases (CompTox) for retention index and ionization efficiency prediction.
• Employed distance metrics and k-nearest neighbor regression for structural similarity and measurability predictions.

Main Results:

• The framework successfully predicts chemical neighbors amenable to detection and estimates method-specific coverage.
• Identified shared and unique measurable chemical regions across different experimental setups.
• Demonstrated that orthogonal methods can expand chemical coverage and diversity in NTA.

Conclusions:

• The developed in silico framework provides a scalable strategy for predicting detectable chemicals.
• Offers a better understanding of chemical measurability within the context of NTA.
• Facilitates improved experimental design for comprehensive exposomic and environmental studies.