Intelligent differential ion mobility spectrometry (iDMS): A deep neural network that predicts optimal space-resolved ion mobility parameters for isomeric monoglycosphingolipids

Monoglycosphingolipids, a class of lipids found in cells, can be in various stereoisomeric forms that are challenging to detect and quantify. Standard techniques like LC-ESI-MS/MS fall short in resolving beta-glucosyl and beta-galactosyl epimers. The introduction of field asymmetric ion mobility spectrometry (FAIMS) or differential mobility mass spectrometry (DMS) as an orthogonal separation technique enables the separation of these epimeric ion clusters based on their response to high and low electric fields.

However, determining the optimal separation voltage (SV) and compensation voltage (CoV) for each lipid is a time-consuming and labor-intensive process that typically relies on pure synthetic standards. To streamline this process, researchers have developed an intelligent DMS (iDMS) system, an in silico supervised neural network model that learns the relationships between SV, CoV, and the structural features of monoglycosphingolipids, such as the sugar headgroup, N-acyl chain length, and N-acyl degree of unsaturation.

By training on a dataset of measured signal intensities from 12 lipids, iDMS can predict the SV and CoV combinations required to resolve any stereoisomer pair. This machine learning approach promises to revolutionize the routine and rapid quantification of biologically relevant monoglycosphingolipid stereoisomers, accelerating the deployment of multiple-reaction-monitoring mode (MRM) RPLC-ESI-DMS-MS/MS assays.