Dataset Information

Robust data-driven incorporation of prior knowledge into the inference of dynamic regulatory networks.

ABSTRACT: Inferring global regulatory networks (GRNs) from genome-wide data is a computational challenge central to the field of systems biology. Although the primary data currently used to infer GRNs consist of gene expression and proteomics measurements, there is a growing abundance of alternate data types that can reveal regulatory interactions, e.g. ChIP-Chip, literature-derived interactions, protein-protein interactions. GRN inference requires the development of integrative methods capable of using these alternate data as priors on the GRN structure. Each source of structure priors has its unique biases and inherent potential errors; thus, GRN methods using these data must be robust to noisy inputs.We developed two methods for incorporating structure priors into GRN inference. Both methods [Modified Elastic Net (MEN) and Bayesian Best Subset Regression (BBSR)] extend the previously described Inferelator framework, enabling the use of prior information. We test our methods on one synthetic and two bacterial datasets, and show that both MEN and BBSR infer accurate GRNs even when the structure prior used has significant amounts of error (>90% erroneous interactions). We find that BBSR outperforms MEN at inferring GRNs from expression data and noisy structure priors.Code, datasets and networks presented in this article are available at http://bonneaulab.bio.nyu.edu/software.html.Supplementary data are available at Bioinformatics online.

SUBMITTER: Greenfield A

PROVIDER: S-EPMC3624811 | biostudies-other | 2013 Apr

REPOSITORIES: biostudies-other

ACCESS DATA

Publications

Robust data-driven incorporation of prior knowledge into the inference of dynamic regulatory networks.

Greenfield Alex A Hafemeister Christoph C Bonneau Richard R

Bioinformatics (Oxford, England) 20130321 8

<h4>Motivation</h4>Inferring global regulatory networks (GRNs) from genome-wide data is a computational challenge central to the field of systems biology. Although the primary data currently used to infer GRNs consist of gene expression and proteomics measurements, there is a growing abundance of alternate data types that can reveal regulatory interactions, e.g. ChIP-Chip, literature-derived interactions, protein-protein interactions. GRN inference requires the development of integrative methods ...[more]

PMID: 23525069

Similar Datasets

Project description:BACKGROUND:In biomarker discovery, applying domain knowledge is an effective approach to eliminating false positive features, prioritizing functionally impactful markers and facilitating the interpretation of predictive signatures. Several computational methods have been developed that formulate the knowledge-based biomarker discovery as a feature selection problem guided by prior information. These methods often require that prior information is encoded as a single score and the algorithms are optimized for biological knowledge of a specific type. However, in practice, domain knowledge from diverse resources can provide complementary information. But no current methods can integrate heterogeneous prior information for biomarker discovery. To address this problem, we developed the Know-GRRF (know-guided regularized random forest) method that enables dynamic incorporation of domain knowledge from multiple disciplines to guide feature selection. RESULTS:Know-GRRF embeds domain knowledge in a regularized random forest framework. It combines prior information from multiple domains in a linear model to derive a composite score, which, together with other tuning parameters, controls the regularization of the random forests model. Know-GRRF concurrently optimizes the weight given to each type of domain knowledge and other tuning parameters to minimize the AIC of out-of-bag predictions. The objective is to select a compact feature subset that has a high discriminative power and strong functional relevance to the biological phenotype. Via rigorous simulations, we show that Know-GRRF guided by multiple-domain prior information outperforms feature selection methods guided by single-domain prior information or no prior information. We then applied Known-GRRF to a real-world study to identify prognostic biomarkers of prostate cancers. We evaluated the combination of cancer-related gene annotations, evolutionary conservation and pre-computed statistical scores as the prior knowledge to assemble a panel of biomarkers. We discovered a compact set of biomarkers with significant improvements on prediction accuracies. CONCLUSIONS:Know-GRRF is a powerful novel method to incorporate knowledge from multiple domains for feature selection. It has a broad range of applications in biomarker discoveries. We implemented this method and released a KnowGRRF package in the R/CRAN archive.

Project description:Under both physiological and pathological conditions gene expression programs are shaped through the interplay of regulatory proteins and their gene targets, interactions between which form intricate gene regulatory networks (GRN). While the assessment of genome-wide expression for the complete set of genes at a given condition has become rather straight-forward and is performed routinely, we are still far from being able to infer the topology of gene regulation simply by analyzing its "descendant" expression profile. In this work we are trying to overcome the existing limitations for the inference and study of such regulatory networks. We are combining our approach with state-of-the-art gene set enrichment analyses in order to create a tool, called Regulatory Network Enrichment Analysis (RNEA) that will prioritize regulatory and functional characteristics of a genome-wide expression experiment.RNEA combines prior knowledge, originating from manual literature curation and small-scale experimental data, to construct a reference network of interactions and then uses enrichment analysis coupled with a two-level hierarchical parsing of the network, to infer the most relevant subnetwork for a given experimental setting. It is implemented as an R package, currently supporting human and mouse datasets and was herein tested on one test case for each of the two organisms. In both cases, RNEA's gene set enrichment analysis was comparable to state-of-the-art methodologies. Moreover, through its distinguishing feature of regulatory subnetwork reconstruction, RNEA was able to define the key transcriptional regulators for the studied systems as supported from the literature.RNEA constitutes a novel computational approach to obtain regulatory interactions directly from a genome-wide expression profile. Its simple implementation, with minimal requirements from the user is coupled with easy-to-parse enrichment lists and a subnetwork file that may be readily visualized to reveal the most important components of the regulatory hierarchy. The combination of prior information and novel concept of a hierarchical reconstruction of regulatory interactions makes RNEA a very useful tool for a first-level interpretation of gene expression profiles.

Project description:BackgroundThe cellular signaling pathway (network) is one of the main topics of organismic investigations. The intracellular interactions between genes in a signaling pathway are considered as the foundation of functional genomics. Thus, what genes and how much they influence each other through transcriptional binding or physical interactions are essential problems. Under the synchronous measures of gene expression via a microarray chip, an amount of dynamic information is embedded and remains to be discovered. Using a systematically dynamic modeling approach, we explore the causal relationship among genes in cellular signaling pathways from the system biology approach.ResultsIn this study, a second-order dynamic model is developed to describe the regulatory mechanism of a target gene from the upstream causality point of view. From the expression profile and dynamic model of a target gene, we can estimate its upstream regulatory function. According to this upstream regulatory function, we would deduce the upstream regulatory genes with their regulatory abilities and activation delays, and then link up a regulatory pathway. Iteratively, these regulatory genes are considered as target genes to trace back their upstream regulatory genes. Then we could construct the regulatory pathway (or network) to the genome wide. In short, we can infer the genetic regulatory pathways from gene-expression profiles quantitatively, which can confirm some doubted paths or seek some unknown paths in a regulatory pathway (network). Finally, the proposed approach is validated by randomly reshuffling the time order of microarray data.ConclusionWe focus our algorithm on the inference of regulatory abilities of the identified causal genes, and how much delay before they regulate the downstream genes. With this information, a regulatory pathway would be built up using microarray data. In the present study, two signaling pathways, i.e. circadian regulatory pathway in Arabidopsis thaliana and metabolic shift pathway from fermentation to respiration in yeast Saccharomyces cerevisiae, are reconstructed using microarray data to evaluate the performance of our proposed method. In the circadian regulatory pathway, we identified mainly the interactions between the biological clock and the photoperiodic genes consistent with the known regulatory mechanisms. We also discovered the now less-known regulations between crytochrome and phytochrome. In the metabolic shift pathway, the casual relationship of enzymatic genes could be detected properly.

Dataset Information

Robust data-driven incorporation of prior knowledge into the inference of dynamic regulatory networks.

Publications

Robust data-driven incorporation of prior knowledge into the inference of dynamic regulatory networks.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure