Advancing the large-scale CCS database for metabolomics and lipidomics at the machine-learning era

Area of Science:

• Analytical Chemistry
• Biochemistry
• Computational Biology

Background:

• Metabolomics and lipidomics aim to comprehensively measure dynamic changes in biological metabolites and lipids.
• Ion mobility-mass spectrometry (IM-MS) aids in separating and identifying these molecules in complex samples.
• Collision cross-section (CCS) values from IM-MS are crucial for unambiguous identification.

Purpose of the Study:

• To review machine learning-based prediction approaches for generating large-scale CCS databases.
• To highlight the utility of CCS databases in supporting metabolomics and lipidomics applications.

Main Methods:

• Discusses recently developed machine learning-based prediction approaches.
• Focuses on the efficient generation of precise CCS databases at scale.

Main Results:

• Machine learning enables large-scale, precise CCS database generation.
• This overcomes limitations of experimentally measured and computationally modeled CCS values.

Conclusions:

• Machine learning-driven CCS databases significantly advance IM-MS applications in metabolomics and lipidomics.
• These databases facilitate broader and more accurate metabolite and lipid identification.