Classifying Host-Guest Topology with Ion Mobility-Mass Spectrometry and Machine Learning.

Area of Science:

• Supramolecular Chemistry
• Computational Chemistry
• Analytical Chemistry

Background:

• Understanding host-guest complex topology is crucial for designing supramolecular assemblies.
• Data-driven methods are increasingly successful in predicting molecular interactions.

Purpose of the Study:

• To evaluate a combined approach using IMS-MS, DFT featurization, and machine learning.
• To predict and classify binding modes of cucurbit[6]uril (CB6) and diamine guest complexes.

Main Methods:

• Ion mobility-mass spectrometry (IMS-MS) for experimental data.
• Density functional theory (DFT) for molecular descriptor generation.
• Machine learning regression models trained on DFT and IMS-MS data.

Main Results:

• A regression model accurately predicted collisional cross sections (CCS) for various host-guest complexes.
• Predicted CCS values naturally separated into distinct groups, enabling topology classification.
• The approach demonstrated effective capture of host-guest topology.

Conclusions:

• The combined DFT-featurization and IMS-MS approach provides a robust framework for predicting host-guest topology.
• This method facilitates the data-driven design of supramolecular assemblies.
• The study highlights the power of integrating computational and experimental techniques.