Many natural processes consist of networks of interacting elements that, over time, affect each other’s state. Their dynamics depend on the pattern of connections and the updating rules for each element. Genomic regulatory networks are networks of this sort. In this paper we use artificial neural networks as a model of the dynamics of gene expression. The significance of the regulatory effect of one gene product on the expression of other genes of the system is defined by a weight matrix. The model considers multigenic regulation including positive and/or negative feedback. The process of gene expression is described by a single network and by two linked networks where transcription and translation are modeled independently. Each of these processes is described by different network controlled by different weight matrices. Methods for computing the parameters of the model from experimental data are discussed. Results computed by means of the model are compared with experimental observations. Generalization to a ‘black box’ concept, where the molecular processes occurring in the cell are considered as signal processing units forming a global regulatory network, is discussed.— Vohradský, J. Neural network model of gene expression. FASEB J. 15, 846‐854 (2001)

Many cell control processes consist of networks of interacting elements that affect the state of each other over time. Such an arrangement resembles the principles of artificial neural networks, in which the state of a particular node depends on the combination of the states of other neurons. The λ bacteriophage lysis/lysogeny decision circuit can be represented by such a network. It is used here as a model for testing the validity of a neural approach to the analysis of genetic networks. The model considers multigenic regulation including positive and negative feedback. It is used to simulate the dynamics of the lambda phage regulatory system; the results are compared with experimental observation. The comparison proves that the neural network model describes behavior of the system in full agreement with experiments; moreover, it predicts its function in experimentally inaccessible situations and explains the experimental observations. The application of the principles of neural networks to the cell control system leads to conclusions about the stability and redundancy of genetic networks and the cell functionality. Reverse engineering of the biochemical pathways from proteomics and DNA micro array data using the suggested neural network model is discussed.

Microarray studies are capable of providing data for temporal gene expression patterns of thousands of genes simultaneously, comprising rich but cryptic information about transcriptional control. However available methods are still not adequate in extraction of useful information about transcriptional regulation from these data. This study presents a dynamic model of gene expression which allows for identification of transcriptional regulators using time series of gene expression. The algorithm was applied for identification of transcriptional regulators controlling 40 cell cycle regulated genes of Saccharomyces cerevisiae. The presented algorithm uses a dynamic model of time continuous gene expression with the assumption that the target gene expression profile results from the action of the upstream regulator. The goal is to apply the model to putative regulators to estimate the transcription pattern of a target gene using a least squares minimization procedure. The procedure iteratively tests all possible transcription factors and selects those that best approximate the target gene expression profile. Results were compared with independently published data and good agreement between the published and identified transcriptional regulators was found.

The first systematic study of small non-coding RNAs (sRNA, ncRNA) in Streptomyces is presented. Except for a few exceptions, the Streptomyces sRNAs, as well as the sRNAs in other genera of the Actinomyces group, have remained unstudied. This study was based on sequence conservation in intergenic regions of Streptomyces, localization of transcription termination factors, and genomic arrangement of genes flanking the predicted sRNAs. Thirty-two potential sRNAs in Streptomyces were predicted. Of these, expression of 20 was detected by microarrays and RT-PCR. The prediction was validated by a structure based computational approach. Two predicted sRNAs were found to be terminated by transcription termination factors different from the Rho-independent terminators. One predicted sRNA was identified computationally with high probability as a Streptomyces 6S RNA. Out of the 32 predicted sRNAs, 24 were found to be structurally dissimilar from known sRNAs. Streptomyces is the largest genus of Actinomyces, whose sRNAs have not been studied. The Actinomyces is a group of bacterial species with unique genomes and phenotypes. Therefore, in Actinomyces, new unique bacterial sRNAs may be identified. The sequence and structural dissimilarity of the predicted Streptomyces sRNAs demonstrated by this study serve as the first evidence of the uniqueness of Actinomyces sRNAs.

HrdB in streptomycetes is a principal sigma factor whose deletion is lethal. This is also the reason why its regulon has not been investigated so far. To overcome experimental obstacles, for investigating the HrdB regulon, we constructed a strain whose HrdB protein was tagged by an HA epitope. ChIP-seq experiment, done in 3 repeats, identified 2137 protein-coding genes organized in 337 operons, 75 small RNAs, 62 tRNAs, 6 rRNAs and 3 miscellaneous RNAs. Subsequent kinetic modeling of regulation of protein-coding genes with HrdB alone and with a complex of HrdB and a transcriptional cofactor RbpA, using gene expression time series, identified 1694 genes that were under their direct control. When using the HrdB–RbpA complex in the model, an increase of the model fidelity was found for 322 genes. Functional analysis revealed that HrdB controls the majority of gene groups essential for the primary metabolism and the vegetative growth. Particularly, almost all ribosomal protein-coding genes were found in the HrdB regulon. Analysis of promoter binding sites revealed binding motif at the −10 region and suggested the possible role of mono- or di-nucleotides upstream of the −10 element.