Network inference with Granger causality ensembles on single-cell transcriptomics

• Published in: Cell reports (Cambridge) Vol. 38; no. 6; p. 110333
• Main Authors: Deshpande, Atul, Chu, Li-Fang, Stewart, Ron, Gitter, Anthony
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 08.02.2022; Elsevier
• Subjects: Algorithms; Animals; Cell Differentiation - physiology; Computational Biology - methods; Gene Expression Profiling - methods; Gene Regulatory Networks - physiology; Mice; mouse embryonic stem cells; network evaluation; pseudotime; Software; time series analysis; transcriptional regulation; Transcriptome - physiology; time series analysis; mouse embryonic stem cells; pseudotime; transcriptional regulation; network evaluation
• ISSN: 2211-1247; 2211-1247
• DOI: 10.1016/j.celrep.2022.110333

Cellular gene expression changes throughout a dynamic biological process, such as differentiation. Pseudotimes estimate cells’ progress along a dynamic process based on their individual gene expression states. Ordering the expression data by pseudotime provides information about the underlying regulator-gene interactions. Because the pseudotime distribution is not uniform, many standard mathematical methods are inapplicable for analyzing the ordered gene expression states. Here we present single-cell inference of networks using Granger ensembles (SINGE), an algorithm for gene regulatory network inference from ordered single-cell gene expression data. SINGE uses kernel-based Granger causality regression to smooth irregular pseudotimes and missing expression values. It aggregates predictions from an ensemble of regression analyses to compile a ranked list of candidate interactions between transcriptional regulators and target genes. In two mouse embryonic stem cell differentiation datasets, SINGE outperforms other contemporary algorithms. However, a more detailed examination reveals caveats about poor performance for individual regulators and uninformative pseudotimes. [Display omitted] •Pseudotime estimates order cells in a dynamic process using single-cell gene expression•SINGE infers gene regulatory networks from gene expression trends over pseudotime•SINGE’s ensembling considers many smoothed versions of irregular pseudotemporal data•Uninformative pseudotime values can be detrimental to network reconstruction Deshpande et al. present SINGE, an algorithm to infer gene regulatory networks from ordered single-cell gene expression data. SINGE uses kernel-based regression to smooth noisy, ordered single-cell data and ensembling to prioritize reliable regulatory relationships.