Constructing Gene Networks Using Variational Bayesian Variable Selection

Abstract

We propose a Bayesian approach for constructing gene networks based on microarray data. Especially, we focus on Bayesian methods that can provide soft (probabilistic) information. This soft information is attractive not only for its ability to measure the level of confidence of the solution, but also because it can be used to realize Bayesian data integration, an extremely important task in gene network research. We propose a variable selection formulation of gene regulation and develop an inference solution based on a variational Bayesian expectation maximization (VBEM) learning rule. This solution has better performance and lower complexity than the popular Monte Carlo sampling techniques. In addition, we develop a method to incorporate the often needed constraints into the VBEM algorithm, making it much more suitable for common cases of small data size. To further illustrate the advantage of the VBEM algorithm, we demonstrate a Bayesian data integration scheme using the soft information obtained from the VBEM algorithm. The efficacy of the proposed VBEM algorithm and the corresponding Bayesian data integration scheme is evaluated on both simulated data and the yeast cell cycle microarray data sets.

Inferring the skeleton cell cycle regulatory network of malaria parasite using comparative genomic and variational Bayesian approaches

“…Also, the VBEM algorithm can report the APPs of the learned network, providing a measurement of confidence on the inference results. Previously, we have tested the proposed algorithm on the time series microarray data from Saccharomyces cerevisiae, the baker's yeast; the inferred topology and regulatory relationship in the cell cycle network reflected experimental evidence (Tienda-Luna et al 2007). In this study, we extend the VBEM learning rule to the malaria parasite.…”

“…Systems biology views of the Plasmodium parasite will take the form of network models. Such models place the modeled system components in a network of relationships that explains the observed data and takes the uncertainty surrounding the relationships into account (Friedman 2004;Kitano 2002;Lilburn and Wang 2006;Segal et al 2005;Tienda-Luna et al 2007). An ability to see and describe the systems-level processes in the organism and thereby identify potential vulnerabilities could result in new malarial control strategies, as shown by pilot studies in parasite genomics, transcriptomics, proteomics, and metabolomics (Aravind et al 2003;Date and Stoeckert 2006;Hall et al 2005;Llinas et al 2006;Yeh et al 2004).…”

Reconstructing time series GRN using a neuro-fuzzy system

“…Gene networks have been modeled according to various approaches [3,4,12]. Although there have been many proposed algorithms for reconstructing gene regulatory networks, each algorithm has specific disadvantages during inference of the gene regulatory network.…”

“…For example, the dynamic Bayesian network model [12] based on time series data constructs a gene network with cyclic regulatory information, necessitating that the data be discretized into several classes; results depend on the discretization thresholds, leading to information loss. The model based on the Variational Bayes Expectation Maximization (VBEM) algorithm [3] regulatory minimization [2,14]. The time delay algorithm for the reconstruction of an accurate cellular network (ARACNE) results in low reconstruction accuracy.…”