Machine learning approaches offer the potential to systematically identify transcriptional regulatory interactions from a compendium of microarray expression profiles. However, experimental validation of the performance of these methods at the genome scale has remained elusive. Here we assess the global performance of four existing classes of inference algorithms using 445 Escherichia coli Affymetrix arrays and 3,216 known E. coli regulatory interactions from RegulonDB. We also developed and applied the context likelihood of relatedness (CLR) algorithm, a novel extension of the relevance networks class of algorithms. CLR demonstrates an average precision gain of 36% relative to the next-best performing algorithm. At a 60% true positive rate, CLR identifies 1,079 regulatory interactions, of which 338 were in the previously known network and 741 were novel predictions. We tested the predicted interactions for three transcription factors with chromatin immunoprecipitation, confirming 21 novel interactions and verifying our RegulonDB-based performance estimates. CLR also identified a regulatory link providing central metabolic control of iron transport, which we confirmed with real-time quantitative PCR. The compendium of expression data compiled in this study, coupled with RegulonDB, provides a valuable model system for further improvement of network inference algorithms using experimental data.

/pdf/large-scale-mapping-and-validation-of-escherichia-coli-5efq1yb3s5.pdf

Large-scale mapping and validation of Escherichia coli transcriptional regulation from a compendium of expression profiles.

A major goal of biology is to provide a quantitative description of cellular behaviour. This task, however, has been hampered by the difficulty in measuring protein abundances and their variation. Here we present a strategy that pairs high-throughput flow cytometry and a library of GFP-tagged yeast strains to monitor rapidly and precisely protein levels at single-cell resolution. Bulk protein abundance measurements of >2,500 proteins in rich and minimal media provide a detailed view of the cellular response to these conditions, and capture many changes not observed by DNA microarray analyses. Our single-cell data argue that noise in protein expression is dominated by the stochastic production/destruction of messenger RNAs. Beyond this global trend, there are dramatic protein-specific differences in noise that are strongly correlated with a protein's mode of transcription and its function. For example, proteins that respond to environmental changes are noisy whereas those involved in protein synthesis are quiet. Thus, these studies reveal a remarkable structure to biological noise and suggest that protein noise levels have been selected to reflect the costs and potential benefits of this variation.

Single-cell proteomic analysis of S. cerevisiae reveals the architecture of biological noise

Genome-wide map of nucleosome acetylation and methylation in yeast.

Reconstructing gene regulatory networks from high-throughput data is a long-standing challenge. Through the Dialogue on Reverse Engineering Assessment and Methods (DREAM) project, we performed a comprehensive blind assessment of over 30 network inference methods on Escherichia coli, Staphylococcus aureus, Saccharomyces cerevisiae and in silico microarray data. We characterize the performance, data requirements and inherent biases of different inference approaches, and we provide guidelines for algorithm application and development. We observed that no single inference method performs optimally across all data sets. In contrast, integration of predictions from multiple inference methods shows robust and high performance across diverse data sets. We thereby constructed high-confidence networks for E. coli and S. aureus, each comprising ~1,700 transcriptional interactions at a precision of ~50%. We experimentally tested 53 previously unobserved regulatory interactions in E. coli, of which 23 (43%) were supported. Our results establish community-based methods as a powerful and robust tool for the inference of transcriptional gene regulatory networks.

/pdf/wisdom-of-crowds-for-robust-gene-network-inference-3w97eu1eis.pdf

Wisdom of crowds for robust gene network inference

Eukaryotic genomes are packaged into nucleosome particles that occlude the DNA from interacting with most DNA binding proteins. Nucleosomes have higher affinity for particular DNA sequences, reflecting the ability of the sequence to bend sharply, as required by the nucleosome structure. However, it is not known whether these sequence preferences have a significant influence on nucleosome position in vivo, and thus regulate the access of other proteins to DNA. Here we isolated nucleosome-bound sequences at high resolution from yeast and used these sequences in a new computational approach to construct and validate experimentally a nucleosome–DNA interaction model, and to predict the genome-wide organization of nucleosomes. Our results demonstrate that genomes encode an intrinsic nucleosome organization and that this intrinsic organization can explain ∼50% of the in vivo nucleosome positions. This nucleosome positioning code may facilitate specific chromosome functions including transcription factor binding, transcription initiation, and even remodelling of the nucleosomes themselves.

http://absimage.aps.org/image/MAR08/MWS_MAR08-2008-020581.pdf

The Genomic Code for Nucleosome Positioning

DNA-binding transcriptional regulators interpret the genome's regulatory code by binding to specific sequences to induce or repress gene expression. Comparative genomics has recently been used to identify potential cis-regulatory sequences within the yeast genome on the basis of phylogenetic conservation, but this information alone does not reveal if or when transcriptional regulators occupy these binding sites. We have constructed an initial map of yeast's transcriptional regulatory code by identifying the sequence elements that are bound by regulators under various conditions and that are conserved among Saccharomyces species. The organization of regulatory elements in promoters and the environment-dependent use of these elements by regulators are discussed. We find that environment-specific use of regulatory elements predicts mechanistic models for the function of a large population of yeast's transcriptional regulators.

/pdf/transcriptional-regulatory-code-of-a-eukaryotic-genome-2df92bvrme.pdf

Transcriptional regulatory code of a eukaryotic genome

We discuss a method of combining genome-wide transcription factor binding data, gene expression data, and genome sequence data for the purpose of motif discovery in S. cerevisiae. Within the word-counting algorithmic approach to motif discovery, we present a method of incorporating information from negative intergenic regions where a transcription factor is thought not to bind, and a statistical significance measure which account for intergenic regions of different lengths. Our results demonstrate that our method performs slightly better than other motif discovery algorithms. Finally, we present significant potential new motifs discovered by the algorithm.

/pdf/negative-information-for-motif-discovery-2kxp9231vi.pdf

Ken T. Takusagawa

Papers

Transcriptional regulatory code of a eukaryotic genome

Negative information for motif discovery.