A Bayesian framework for the inference of gene regulatory networks from time and pseudo-time series data

• Published in: Bioinformatics Vol. 34; no. 6; pp. 964 - 970
• Main Authors: Sanchez-Castillo, M, Blanco, D, Tienda-Luna, I M, Carrion, M C, Huang, Yufei
• Format: Journal Article
• Language: English
• Published: England Oxford University Press 15.03.2018
• Subjects: Algorithms; Animals; Bayes Theorem; Computational Biology - methods; Gene Expression Profiling - methods; Gene Regulatory Networks; Mice; Models, Biological; Sequence Analysis, RNA - methods; Single-Cell Analysis - methods
• ISSN: 1367-4803; 1460-2059; 1367-4811
• DOI: 10.1093/bioinformatics/btx605

Abstract Motivation Molecular profiling techniques have evolved to single-cell assays, where dense molecular profiles are screened simultaneously for each cell in a population. High-throughput single-cell experiments from a heterogeneous population of cells can be experimentally and computationally sorted as a sequence of samples pseudo-temporally ordered samples. The analysis of these datasets, comprising a large number of samples, has the potential to uncover the dynamics of the underlying regulatory programmes. Results We present a novel approach for modelling and inferring gene regulatory networks from high-throughput time series and pseudo-temporally sorted single-cell data. Our method is based on a first-order autoregressive moving-average model and it infers the gene regulatory network within a variational Bayesian framework. We validate our method with synthetic data and we apply it to single cell qPCR and RNA-Seq data for mouse embryonic cells and hematopoietic cells in zebra fish. Availability and implementation The method presented in this article is available at https://github.com/mscastillo/GRNVBEM.