A random graph generation model for transcription networks and nonparametric simulator for RNA-seq expression data

Abstract

With the advent of high-throughput sequencing and continually decreasing costs, an abundance of gene expression data has been generated and is available for analysis. These data provide a rich resource for inferring connections in gene regulatory networks. A vast collection of methodologies have been developed, but a challenge remains in assessing and benchmarking their performance. Gold-standard datasets are scarce, and commonly-used simulators are not designed to resemble the data generated from RNA-seq experiments. The present study provides a novel random graph generator that produces networks having comparable topology to transcription networks. In addition, a nonparametric simulator is proposed that generates conditionally dependent expression data; the conditional dependencies are based on an underlying network structure, and the marginal distribution of gene expression profiles are based on a reference RNA-seq dataset. These methods provide tools for creating in silico RNA-seq data for benchmarking and assessing gene network inference methods. Keywords: data science, random graph, transcription network, RNA expression