Automated supervised learning pipeline for non-targeted GC-MS data analysis.

Area of Science:

• Analytical Chemistry
• Chemometrics
• Data Science

Background:

• Non-targeted analysis is crucial in metabolomics, environmental, and food analysis.
• Conventional gas chromatography-mass spectrometry (GC-MS) data processing involves error-prone steps like baseline correction, feature detection, and retention time alignment.
• These steps often necessitate time-consuming manual corrections.

Purpose of the Study:

• To introduce a novel, fully automated approach for non-targeted GC-MS data processing.
• To eliminate the need for feature extraction and retention time alignment.
• To leverage supervised machine learning for efficient data analysis.

Main Methods:

• Utilizes decomposed tensors of segmented chromatographic raw data signals.
• Applies supervised machine learning to rank chromatogram regions.
• Focuses on identifying regions that differentiate between sample classes.

Main Results:

• Demonstrates a fully automated data processing workflow for non-targeted GC-MS.
• Successfully ranks relevant chromatographic regions using machine learning.
• Validates the approach's performance on three published datasets.

Conclusions:

• The novel automated approach offers a significant improvement over conventional GC-MS data processing.
• It reduces manual intervention and potential errors, saving time and resources.
• This method enhances the efficiency and reliability of non-targeted analysis in various chemical domains.