Comparison and Optimization of Cellular Neighbor Preference Methods for Quantitative Tissue Analysis

Abstract

Studying the spatial distribution of cell types in tissues is essential for understanding their function in health and disease. A widely used spatial feature for quantifying tissue organization is the pairwise neighbor preference (NEP) of cell types, commonly referred to as co-occurrence or colocalization. Various methods to infer NEP have proved their utility in spatial omics studies, but despite their broad usage, no clear guidelines exist for selecting one method over the other. In this paper, we deconstruct frequently used methods into their underlying analysis steps and evaluate their optimal combination. We studied the methods on two aspects: (1) their discriminatory power to distinguish different tissue architectures and (2) their ability to recover the directionality of NEPs. We compared existing as well as our in-house developed method (conditional z-score (COZI)) and compared their performance using in silico tissue simulations and demonstrated its biological applicability in a myocardial infarction dataset. Overall, our study serves as a comprehensive guide for users and method developers in spatial omics analysis and offers a novel approach (COZI), which outperforms existing methods, to performing NEP analysis.