SeqNet: An R Package for Generating Gene-Gene Networks and Simulating RNA-Seq Data

Abstract

Gene expression data provide an abundant resource for inferring connections in gene regulatory networks. While methodologies developed for this task have shown success, a challenge remains in comparing the performance among methods. Gold-standard datasets are scarce and limited in use. And while tools for simulating expression data are available, they are not designed to resemble the data obtained from RNA-seq experiments. SeqNet is an R package that provides tools for generating a rich variety of gene network structures and simulating RNA-seq data from them. This produces in silico RNA-seq data for benchmarking and assessing gene network inference methods.

A novel probabilistic generator for large-scale gene association networks

“…performance of network inference methods [5]. As a consequence, the relative performance in these simulation studies may not generalize to real datasets.…”

“…This provides a useful model for generating data from an undirected graph, and it can be used to simulate gene expression data that resemble RNA-seq datasets [5].…”

“…In particular, it serves to replace the probabilistic generators commonly used in simulation studies. The generated association networks can be used in a GGM or other Markov model to simulate gene expression data [5]. Performance estimates from these simulations will more closely reflect the performance on real observational gene expression data.…”

“…However, there are two key aspects that set it apart: (1) the network is constructed by iteratively creating overlapping modular structures, and (2) it uses a more flexible model for preferential attachment. A penultimate version of this algorithm was proposed in [5]. This manuscript serves as a refinement of that algorithm and analyzes its statistical properties.…”

“…The focus of this manuscript is on the first step-generating association networks. Once a network is generated, gene expression data can be simulated as described in [5], for instance. We focus on the first step because the performance of network inference methods is heavily influenced by the topology of the underlying network [4,5], hence it is crucial for simulated networks to resemble real association networks.…”