Influence of Network Topology and Data Collection on Network Inference

Abstract

We recently developed an approach for testing the accuracy of network inference algorithms by applying them to biologically realistic simulations with known network topology. Here, we seek to determine the degree to which the network topology and data sampling regime influence the ability of our Bayesian network inference algorithm, NETWORKINFERENCE, to recover gene regulatory networks. NETWORKINFERENCE performed well at recovering feedback loops and multiple targets of a regulator with small amounts of data, but required more data to recover multiple regulators of a gene. When collecting the same number of data samples at different intervals from the system, the best recovery was produced by sampling intervals long enough such that sampling covered propagation of regulation through the network but not so long such that intervals missed internal dynamics. These results further elucidate the possibilities and limitations of network inference based on biological data.

A framework for integrating the songbird brain

“…The detailed report of the invention of this approach is presented in Smith et al (2002), and the further characterizations and improvements using GeneSimulator and RegulationRecover will be by J. Yu et al (unpublished observations) and BrainSim and NetworkInference is in Smith et al (2003).…”

Importance of Input Perturbations and Stochastic Gene Expression in the Reverse Engineering of Genetic Regulatory Networks: Insights From an Identifiability Analysis of an In Silico Network

“…Wessels et al (2001) explored several approaches for reverse engineering genetic regulatory networks from gene expression data, but they constrained the complexity of their in silico systems by the reverse engineering approaches themselves. The studies of Zak et al (2001a), Smith et al (2002Smith et al ( , 2003, and Yeung et al (2002) more closely paralleled the experimental situation in that their reverse engineering techniques differed from the systems used to generate the simulation data. Whereas Smith et al (2002Smith et al ( , 2003 used largely descriptive in silico models, the in silico models of Zak et al (2001a) and Yeung et al (2002) were based on simplified biochemical models of transcriptional regulation.…”

“…The studies of Zak et al (2001a), Smith et al (2002Smith et al ( , 2003, and Yeung et al (2002) more closely paralleled the experimental situation in that their reverse engineering techniques differed from the systems used to generate the simulation data. Whereas Smith et al (2002Smith et al ( , 2003 used largely descriptive in silico models, the in silico models of Zak et al (2001a) and Yeung et al (2002) were based on simplified biochemical models of transcriptional regulation. Finally, Michaud et al (2003) developed an online tool that allows users to generate artificial data sets from known networks of arbitrary structure, but like Smith et al (2002Smith et al ( , 2003, the models are of a descriptive nature.…”

Fundamental limitations of network reconstruction from temporal data

“…Note also that, if dim ker M i ¼ d i , providing the value of Pða ij Þ for d i entries of a i as prior information guarantees a correct reconstruction of Pða i Þ. Sparsity of the interaction matrix can be equally effective prior information (electronic supplementary material 9). Our analysis lays a firm theoretical basis to some fundamental limitations of network reconstruction that have been observed before via computer simulations-such as the role of the informativeness of the temporal data [34] and the fact that different networks may produce exactly the same temporal response [35]. A particular instance of our analysis is when different networks, such as those obtained from chemical reactions, produce the same dynamical model (i.e.…”

“…However, after a decade of extensive studies, some core problems in NR remain open [27][28][29][30][31][32][33]. In particular, we still lack an identifiability analysis characterizing the conditions on the temporal data and knowledge of the coupling functions that are necessary to reconstruct a desired property PðAÞ [34,35]. Our main contribution here is to derive such necessary conditions in the ideal case (i.e.…”