Hybrid Method Inference for the Construction of Cooperative Regulatory Network in Human

Abstract

Reconstruction of large scale gene regulatory networks (GRNs in the following) is an important step for understanding the complex regulatory mechanisms within the cell. Many modeling approaches have been introduced to find the causal relationship between genes using expression data. However, they have been suffering from high dimensionality-large number of genes but a small number of samples, overfitting, heavy computation time and low interpretability. We have previously proposed an original Data Mining algorithm Licorn, that infers cooperative regulation network from expression datasets. In this work, we present an extension of Licorn to a hybrid inference method h-Licorn that uses search in both discrete and real valued spaces. Licorn's algorithm, using the discrete space to find cooperative regulation relationships fitting the target gene expression, has been shown to be powerful in identifying cooperative regulation relationships that are out of the scope of most GRN inference methods. Still, as many of related GRN inference techniques, Licorn suffers from a large number of false positives. We propose here an extension of Licorn with a numerical selection step, expressed as a linear regression problem, that effectively complements the discrete search of Licorn. We evaluate a bootstrapped version of h-Licorn on the in silico Dream5 dataset and show that h-Licorn has significantly higher performance than Licorn, and is competitive or outperforms state of the art GRN inference algorithms, especially when operating on small data sets. We also applied h-Licorn on a real dataset of human bladder cancer and show that it performs better than other methods in finding candidate regulatory interactions. In particular, solely based on gene expression data, h-Licorn is able to identify experimentally validated regulator cooperative relationships involved in cancer.

Time-course transcriptome analysis of human cellular reprogramming from multiple cell types reveals the drastic change occurs between the mid phase and the late phase

“…The CoRegNet infers cooperative TF network and scores TF influences with the h-LICORN algorithm by using TFs and target genes expression profiles (Fig. 1b ) [ 34 ]. To reconstruct regulatory networks, we set the parameter of minCoregSupport as 0.55 due to the limitation on computational memory, where the parameter indicates how frequently the set of co-regulators appears in the dataset (Fig.…”

Latent network-based representations for large-scale gene expression data analysis

“…We propose to consider a GRN as a background network structure that defines the relations between genes (used as features in gene expression-based data analyses) and to exploit this structure to perform a transformation of the input signal of expression for unravelling latent signals that are more informative than the initial expression data. In this study, we use LICORN [15,18] approach for the inference of regulatory networks. LICORN identifies groups of regulators as co-activators A and co-inhibitors I for each target gene.…”

“…In order to obtain a graph structure that presents the underlying relations between the features (genes) of a transcriptomic dataset, we use LICORN [15,18] (available in the COREGNET Bioconductor R package [3]). LICORN is a data mining algorithm that allows the inference of gene regulatory networks that can capture the targets of transcription factors from genome wide expression data.…”

CoRegNet: reconstruction and integrated analysis of co-regulatory networks

“…To reconstruct a large-scale regulatory network from gene expression data, the C o R eg N et package implements the h -L icorn algorithm ( Chebil et al. , 2014 ; Elati et al.…”

“…Compared to current methods, h -L icorn focuses on the identification of cooperative regulators of genes. It was proven to have comparable TF-gene pairs prediction performance with state of the art methods in synthetic and Human datasets ( Chebil et al. , 2014 ) and to retrieve more plausible cooperative TF pairs in yeast ( Elati et al ., 2007 , Lai et al ., 2014 ) and Human datasets (see Supplementary Information ).…”