PIA SSN: Parallel Image Acquisition and Spatial Similarity Network for Tandem Mass Spectrometry Imaging

Abstract

Unambiguous molecular annotations are essential to discern complex local biochemical processes in spatial biology. Here we present a scalable and broadly applicable platform for tandem mass spectrometry imaging (MS2I) that overcomes current limitations in annotation with MSI by integrating Parallel Image Acquisition (PIA) with a novel open-access computational framework, Spatial Similarity Networking (SSN). The PIA employs parallelized acquisition of untargeted MSI and targeted MS2I data using multiple inclusion lists to ensure spatially consistent and structure-resolved imaging of hundreds of molecular species in a single experiment. For molecular annotation, we have developed the SSN that complements PIA by leveraging spatial correlations among product ions through a graph-based analysis framework to enable confident molecular annotation even within highly complex MS2I datasets. Using this integrated approach, we successfully resolved and annotated 134 phospholipid isomers and isobars from mouse brain tissue and suggest confidence levels for annotation for the MSI community. Furthermore, we applied our platform to interrogate cholesterol metabolism in human multiple sclerosis brain tissue, achieving annotation of six novel brain-related oxysterols and revealing spatially correlated oxidation pathways linked to lesion severity. Together, PIA and SSN establish a new framework for large-scale, structure-specific mass spectrometry imaging, with broad implications for spatial metabolomics, lipidomics, and chemical pathology beyond current capabilities.