Investigation into Small Molecule Isomeric Glucuronide Metabolite Differentiation Using In Silico and Experimental...

Area of Science:

• Analytical Chemistry
• Computational Chemistry
• Pharmacology

Background:

• Identifying isomeric drug metabolites is complex, often requiring multiple analytical techniques like LC-MS and NMR.
• Ion mobility mass spectrometry (IMMS) offers a potential single-solution for rapid metabolite identification, including isomers.
• Automated computational pipelines comparing experimental and predicted collision cross-section (CCS) values can aid structural assignment.

Purpose of the Study:

• To evaluate the utility of IMMS combined with computational prediction for identifying isomeric drug metabolites.
• To compare the accuracy of quantum mechanics (QM) and machine learning (ML) models in predicting CCS values.
• To assess if the differences in CCS values between isomeric pairs are sufficient for unambiguous structural identification.

Main Methods:

• Utilized an ion mobility enabled Q-Tof mass spectrometer to measure CCS values for 28 small molecule glucuronide metabolites (14 isomeric pairs).
• Compared experimental CCS values against predictions from QM and ML *in silico* models.
• Analyzed the differences in CCS values between isomeric pairs to determine the potential for confident structural distinction.

Main Results:

• Both QM and ML models showed good correlation with experimental CCS values.
• QM and ML *in silico* methods produced similar predicted CCS values.
• The QM model better represented the CCS differences between isomers, though only naringenin glucuronides were confidently distinguished.
• The ML model could not confidently distinguish the studied isomer pairs.

Conclusions:

• IMMS coupled with *in silico* CCS prediction is a valuable tool for metabolite identification.
• While promising, current computational models, including QM and ML, have limitations in confidently distinguishing between all isomeric drug metabolites based solely on CCS differences.
• Further refinement of computational approaches and analysis of analyte structures are needed to improve the resolution of isomeric compounds.