From Analysis to Assessment: Machine Learning for Non-Target Screening of Pollutants Using Chromatography Coupled...

Area of Science:

• Environmental Chemistry
• Analytical Chemistry
• Computational Chemistry

Background:

• Environmental chemical diversity outpaces traditional monitoring.
• Non-target screening (NTS) with high-resolution mass spectrometry (HRMS) is crucial for discovering unknown contaminants.
• Liquid or gas chromatography coupled with ion mobility-mass spectrometry (LC/GC-IM-MS) provides enhanced data for NTS.

Purpose of the Study:

• To review machine learning (ML) algorithms and workflows for NTS in LC/GC-(IM-)MS.
• To synthesize recent advances in ML-enabled NTS applications.
• To highlight ML's role in transforming environmental risk assessment.

Main Methods:

• Review of machine learning algorithms (supervised and unsupervised).
• Analysis of representative workflows in LC/GC-(IM-)MS-NTS.
• Synthesis of ML applications across the NTS pipeline.

Main Results:

• ML efficiently processes high-dimensional data from LC/GC-(IM-)MS for NTS.
• ML algorithms streamline labor-intensive NTS processes.
• ML shows strong potential across sample analysis, data acquisition, and risk assessment.

Conclusions:

• Machine learning is a powerful tool for enhancing NTS.
• ML integration is transforming environmental contaminant identification and risk assessment.
• Future advances in ML will further improve NTS efficiency and robustness.