Mantpy: a framework for extracellular matrix analysis in spatial proteomics

Abstract

Spatial proteomics technologies now profile cells and the extracellular matrix (ECM) together in situ. Yet analysis tools remain cell-centric, despite the ECM playing an essential role in health and disease. Here we present Mantpy, a framework that represents the ECM, and its interface with cells, as spatial graphs. Mantpy builds ECM graphs directly from matrix markers and links them with cell graphs for joint cell-ECM analysis, supporting graph statistics, explainable graph deep learning and visualisation. Mantpy extends from a single ECM marker to multiplexed cell-ECM panels. It recovers layered tissue architecture in the intestine from collagen IV alone, quantifies infection-associated ECM remodelling in mouse liver, and relates cells to the surrounding matrix in mouse lung, where cellular context recovers synthetically corrupted matrix labels. Released with ECM-inclusive datasets and interoperating with the scverse ecosystem and pytorch geometric.