Multi-modal analysis and integration of single-cell morphological data

Abstract

Morphology is an essential phenotype in the characterization of cells and their states, as it is closely related to cell function. Recent advances like Patch-seq enable simultaneously profiling the morphology, gene expression, and physiological properties of individual cells. However, computational methods that can summarize the great diversity of complex cell morphologies found in tissues and infer associations with other single-cell data modalities remain scarce. Here we report a computational framework, named CAJAL, for the morphometric and multi-modal analysis of single-cell data. CAJAL uses concepts from metric geometry to accurately build and visualize cell morphology summary spaces, integrate cellular morphologies across technologies, and establish associations between morphological, molecular, and physiological cellular processes. We demonstrate the utility of CAJAL by applying it to published Patch-seq, patch-clamp, serial electron, and two-photon microscopy data, and show that it represents a substantial improvement in functionality, scope, and accuracy with respect to current methods for cell morphometry.