Enhancing gene regulatory network inference through data integration with markov random fields

Abstract

A gene regulatory network links transcription factors to their target genes and represents a map of transcriptional regulation. Much progress has been made in deciphering gene regulatory networks computationally. However, gene regulatory network inference for most eukaryotic organisms remain challenging. To improve the accuracy of gene regulatory network inference and facilitate candidate selection for experimentation, we developed an algorithm called GRACE (Gene Regulatory network inference ACcuracy Enhancement). GRACE exploits biological a priori and heterogeneous data integration to generate high-confidence network predictions for eukaryotic organisms using Markov Random Fields in a semi-supervised fashion. GRACE uses a novel optimization scheme to integrate regulatory evidence and biological relevance. It is particularly suited for model learning with sparse regulatory gold standard data. We show GRACE's potential to produce high confidence regulatory networks compared to state of the art approaches using Drosophila melanogaster and Arabidopsis thaliana data. In an A. thaliana developmental gene regulatory network, GRACE recovers cell cycle related regulatory mechanisms and further hypothesizes several novel regulatory links, including a putative control mechanism of vascular structure formation due to modifications in cell proliferation.

Distinct tissue-specific transcriptional regulation revealed by gene regulatory networks in maize

“…GENIE3 has implementations in Python, Matlab and R languages that are easy for researchers to use. Also, similar methods have been adapted to time-series [ 30 ], single-cell [ 31 ], and integrated [ 32 ] GRNs showing its wide applicability. GENIE3 takes advantage of parallel computing and can generates large networks on a multi-core desktop.…”

Reprogramming of regulatory network using expression uncovers sex-specific gene regulation in Drosophila

“…Methods to infer GRNs typically combine computational approaches and experimental data collected from different sample types, different conditions, different techniques, and different labs. Such data integration leverages dependencies that can be confidently uncovered thanks to the multitude of surveyed conditions, but leads to context-agnostic wiring diagrams 1 – 3 . These context-agnostic networks do not accommodate regulatory program reality, which is specific to tissue types, developmental stages, sex, and other factors.…”

Statistical and Machine Learning Approaches to Predict Gene Regulatory Networks From Transcriptome Datasets

“…They identified 968 candidates of regulators and represented a subnetwork for a regulatory gene FAC1 by superimposing information from its protein–protein interaction (PPI) network and the differentially expressed genes of its mutant in F. graminearum . GENIE3 6 , a tree-based ML algorithm ( Huynh-Thu et al, 2010 ), has been widely employed in recent GRN inference studies with both static and dynamic transcriptome data from various species ( Banf and Rhee, 2017 ; Desai et al, 2017 ; Redekar et al, 2017 ). Huang et al (2018) applied GENIE3 to infer GRNs with over 1,000 publicly available RNA-Seq data from various tissues such as leaf, root, shoot, apical meristem, and seed, and created four tissue-specific GRNs.…”

“… Redekar et al (2017) used five different algorithms, ARACNE 8 ( Margolin et al, 2006 ), GENIE3 ( Huynh-Thu et al, 2010 ), TIGRESS ( Haury et al, 2012 ), partial correlation (GeneTS 9 ) ( Schafer and Strimmer, 2005 ), and CLR ( Faith et al, 2007 ), to infer the GRNs between TFs and co-expressed modules for seed development in soybean 10 , based on 60 RNA-Seq datasets (three biological replicates, five stages of developing seeds, and four experimental lines), and evaluated the resultant GRNs by comparative analysis with published GRNs of Arabidopsis ( Redekar et al, 2017 ) 10 . Banf and Rhee developed a novel GRN inference strategy called GRACE (Gene Regulatory network inference ACcuracy Enhancement 11 ), which generates GRNs through multiple steps to integrate various knowledge related to the regulation of gene expression: initial network prediction from gene expression data using a random forest regression model and integrating information related to gene regulation, subsequent network module extraction by meta-network construction based on information of functionally related genes, and further selection of regulatory links using ensembles of Markov Random Fields ( Banf and Rhee, 2017 ). To infer the developmental GRN in Arabidopsis , the authors incorporated conserved sequence information in its promoter regions and experimentally determined cis -motifs for TFs, together with gene expression data from 83 tissues and stages, and obtained an initial GRN containing 325 regulators, 4,305 targets, and 10,098 links.…”