scCOSMIX: A Mixed‐Effects Framework for Differential Coexpression and Transcriptional Interactions Modeling in Single‐Cell RNA‐Seq

• Published in: Statistics in medicine Vol. 44; no. 18-19; pp. e70213 - n/a
• Main Authors: Bussing, Anderson, Marra, Giampiero, Fan, Daping, Shinohara, Russell, Tu, Danni, Ho, Yen‐Yi
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.08.2025; Wiley Subscription Services, Inc
• Subjects: Computer Simulation; Differential co‐expression; Gene Expression Profiling - methods; hierarchical study design; Humans; mixed effects; Models, Genetic; Models, Statistical; RNA-Seq - methods; Sequence Analysis, RNA - methods; Single-Cell Analysis - methods; Single-Cell Gene Expression Analysis; single‐cell RNA‐seq; Transcription, Genetic; zero‐inflated copula model; single‐cell RNA‐seq; mixed effects; zero‐inflated copula model; Differential co‐expression; hierarchical study design
• ISSN: 0277-6715; 1097-0258; 1097-0258
• DOI: 10.1002/sim.70213

ABSTRACT Advancements in single‐cell RNA‐sequencing (scRNA‐seq) technologies generate a wealth of gene expression data that provide exciting opportunities for studying gene‐gene interactions systematically at individual cell resolution. Genetic interactions within a cell are tightly regulated and often highly dynamic in response to internal cellular signals and external stimuli. Evidence of these dynamic interactions can often be observed in scRNA‐seq data by examining conditional co‐expression changes. Existing approaches for studying these dynamic interaction changes in scRNA‐seq data do not address the multi‐subject hierarchical design commonly considered in single‐cell experiments. In this paper, we propose a Mixed‐effects framework for differential Coexpression and transcriptional interaction modeling in Single‐Cell RNA‐seq (scCOSMiX) to account for the cell‐cell correlation from the same individual. The proposed copula‐based approach allows the zero‐inflation, marginal, and association parameters to be modeled as functions of covariates with subject‐level random effects, to enable analyses to be tailored to the data under consideration. A series of simulation analyses were conducted to evaluate and compare the performance of scCOSMiX to other existing approaches. We applied the proposed method to both droplet and plate‐based scRNA‐seq data sets GSE266919 and GSE108989 to illustrate its applicability across distinct scRNA‐seq experimental protocols.