Search

Gene expression regulation is an intricate, dynamic phenomenon essential for all biolog ical functions. The necessary instructions for gen expression are encoded in cis-regulatory elements that work together and interact with the RNA polymerase to confer specific spatial and temporal patterns of transcrip tion. Therefore, the identification of these el ements is currently an active area of research in computational analysis of regulatory se quences. However, the problem is difficult since the combinatorial interactions between the regulating factors can be very complex. Here we present a web server, Cis-cop, that identifies cis-regulatory modules given a set of transcription factor binding sites and, ad ditionally, also RNA pol sites for a group of genes.

Genomes of many organisms have been sequenced over the last few years. However, transforming such raw sequence data into knowledge remains a hard task. A great number of prediction programs have been developed to address part of this problem: the location of genes along a genome and their expression. We propose a multi-objective methodology to combine state-of-the-art algorithms into an aggregation scheme in order to obtain optimal methods’ aggregations. The results obtained show a major improvement in sensitivity when our methodology is compared to the performance of individual methods for gene finding and gene expression problems. The methodology proposed here is an automatic method generator, and a step forward to exploit all already existing methods, by providing alternative optimal methods’ aggregations to answer concrete queries for a certain biological problem with a maximized accuracy of the prediction. As more approaches are integrated for each of the presented problems, de novo accuracy can be expected to improve further.

Current tools and techniques devoted to examine the content of large databases are often hampered by their inability to support searches based on criteria that are meaningful to their users. These shortcomings are particularly evident in data banks storing representations of structural data such as biological networks. Conceptual clustering techniques have demonstrated to be appropriate for uncovering relationships between features that characterize objects in structural data. However, typical con ceptual clustering approaches normally recover the most obvious relations, but fail to discover the lessfrequent but more informative underlying data associations. The combination of evolutionary algorithms with multiobjective and multimodal optimization techniques constitutes a suitable tool for solving this problem. We propose a novel conceptual clustering methodology termed evolutionary multiobjective conceptual clustering (EMO-CC), re lying on the NSGA-II multiobjective (MO) genetic algorithm. We apply this methodology to identify conceptual models in struc tural databases generated from gene ontologies. These models can explain and predict phenotypes in the immunoinflammatory response problem, similar to those provided by gene expression or other genetic markers. The analysis of these results reveals that our approach uncovers cohesive clusters, even those comprising a small number of observations explained by several features, which allows describing objects and their interactions from different perspectives and at different levels of detail.

Biomedical research has been revolutionized by high throughput techniques and the enormous amount of data they are able to generate. In particular technology has the capacity to monitor changes in RNA abundance for thou sands of genes simultaneously. The interest shown over microarray analysis methods has rapidly raised. Clustering is widely used in the analysis of microarray data to group genes of interest targeted from microarray experiments on the basis of similarity of expression patterns. In this work we apply two clustering algorithms, K-means and Expecta tion Maximization to particular a problem and we compare the groupings obtained on the basis of the cohesiveness of the gene products associated to the genes in each cluster

Genetic and genomic approaches have been used successfully to assign genes to distinct regulatory networks, but the uncertainty concerning the connec tions between genes, the ambiguity inherent to the biological processes, and the impossibility of experimentally determining the underlying biological properties only allow a rough prediction of the dynamics of genes. Here we describe the GE NIE methodology that formulates alternative models of genetic regulatory networks based on the available literature and transcription factor binding site evidence. It also provides a framework for the analysis of these models optimized by genetic algorithms, inferring their optimal parameters, simulating their behavior, evaluat ing them by integrating robustness, realness and flexibility criteria, and contrasting the predictions to experimentally results obtained by Gene Fluorescence Protein analysis. The application of this method to the regulatory network of the bacterium Salmonella enterica uncovered new mechanisms that enable the inter-connection of the PhoP/PhoQ and the PmrA/PmrB two component systems. The predictions were experimentally verified to establish that both transcriptional and post-transcriptional mechanisms are employed to connect these two systems.

One of the biggest challenges in genomics is the elucidation of the design principles controlling gene expression. Current approaches examine promoter sequences for particular features, such as the presence of binding sites for a transcriptional regulator, and identify recurrent relationships among these features termed network motifs. To define the expression dynamics of a group of genes, the strength of the connections in a network must be specified, and these are determined by the cis-promoter features participating in the regulation. Approaches that homogenize features among promoters (e.g., relying on consensuses to describe the various promoter features) and even across species hamper the discovery of the key differences that distinguish promoters that are co-regulated by the same transcriptional regulator. Thus, we have developed a an approach based on fuzzy logic expressions to analyze proteobacterial genomes for promoter features that is specifically designed to account for the variability in sequence, location and topology intrinsic to differential gene expression. We applied our method to characterize network motifs controlled by the PhoP/PhoQ regulatory system of Escherichia coli and Salmonella enterica serovar Typhimurium. We identify key features that enable the PhoP protein to produce distinct kinetic patterns in target genes, which could not have been uncovered just by inspecting network motifs.

A critical challenge of the postgenomic era is to understand how genes are differentially regulated in and between genetic networks. The fact that such co-regulated genes may be differentially regulated suggests that subtle differences in the shared cis-acting regulatory elements are likely significant, however it is unknown which of these features increase or reduce expression of genes. In principle, this expression can be measured by microarray experi ments, though they incorporate systematic errors, and moreover produce a lim ited classification (e.g. up/down regulated genes). In this work, we present an unsupervised machine learning method to tackle the complexities governing gene expression, which considers gene expression data as one feature among many. It analyzes features concurrently, recognizes dynamic relations and gen erates profiles, which are groups of promoters sharing common features. The method makes use of multiobjective techniques to evaluate the performance of profiles, and has a multimodal approach to produce alternative descriptions of same expression target. We apply this method to probe the regulatory networks governed by the PhoP/PhoQ two-component system in the enteric bacteria Es cherichia coli and Salmonella enterica. Our analysis uncovered profiles that were experimentally validated, suggesting correlations between promoter regu latory features and gene expression kinetics measured by green fluorescent pro tein (GFP) assays.

Biomedical research has been revolutionized by high-throughput techniques and the enormous amount of biological data they are able to generate. The interest shown over network models and systems biology is rapidly raising. Genetic networks arise as an essential task to mine these data since they explain the function of genes in terms of how they influence other genes. Many modeling approaches have been proposed for building genetic networks up. However, it is not clear what the advantages and disadvantages of each model are. There are several ways to discriminate network building models, being one of the most important whether the data being mined presents a static or dynamic fashion. In this work we compare static and dynamic models over a problem related to the inflammation and the host response to injury. We show how both models provide complementary information and cross-validate the obtained results.

The increased availability of biological databases contain ing representations of complex objects permits access to vast amounts of data. In spite of the recent renewed interest in knowledge-discovery tech niques (or data mining), there is a dearth of data analysis methods in tended to facilitate understanding of the represented objects and related systems by their most representative features and those relationship de rived from these features (i.e., structural data). In this paper we propose a conceptual clustering methodology termed EMO-CC for Evolution ary Multi-Objective Conceptual Clustering that uses multi-objective and multi-modal optimization techniques based on Evolutionary Algorithms that uncover representative substructures from structural databases. Be sides, EMO-CC provides annotations of the uncovered substructures, and based on them, applies an unsupervised classification approach to retrieve new members of previously discovered substructures. We apply EMO-CC to the Gene Ontology database to recover interesting sub structures that describes problems from different points of view and use them to explain inmuno-inflammatory responses measured in terms of gene expression profiles derived from the analysis of longitudinal blood expression profiles of human volunteers treated with intravenous endo toxin compared to placebo.

A critical challenge of the postgenomic era is to understand how genes are differentially regulated even when they belong to a given network. Because the fundamental mechanism controlling gene expression operates at the level of transcription initiation, computational techniques have been devel oped that identify cis-regulatory features and map such features into differential expression patterns. The fact that such co-regulated genes may be differentially regulated suggests that subtle differences in the shared cis-acting regulatory elements are likely significant. Thus, we carry out an exhaustive description of cis-acting regulatory features including the orientation, location and number of binding sites for a regulatory protein, the presence of binding site submotifs, the class and number of RNA polymerase sites, as well as gene expression data, which is treated as one feature among many. These features, derived from dif ferent domain sources, are analyzed concurrently, and dynamic relations are re cognized to generate profiles, which are groups of promoters sharing common features. We apply this method to probe the regulatory networks governed by the PhoP/PhoQ two-component system in the enteric bacteria Escherichia coli and Salmonella enterica. Our analysis uncovered novel members of the PhoP regulon as and the resulting profiles group genes that share underlying biologi cal that characterize the system kinetics. The predictions were experimentally validated to establish that the PhoP protein uses multiple mechanisms to control gene transcription and is a central element in a highly connected network.

The rapid development of methods that select over/under expressed genes from RNA microarray experiments have not yet satisfied the need for tools that identify differential profiles that distinguish between experimental conditions such as time, treatment and phenotype. We evaluate several microarray analysis methods and study their performance, finding that none of the methods alone identifies all observable differential profiles, nor subsumes the results obtained by the other methods. Therefore, we propose a machine learning based methodology that identifies and combines the abilities of microarray analysis methods to recognize differential profiles. We encode the results of this methodology in decision making association rules able to decide which method or method-aggregation is optimal to retrieve a set of genes exhibiting a common profile. These solutions are optimal in the sense that they constitute partial ordered subsets of all method-aggregations bounded by the most specific and the most sensitive available solution. This methodology was successfully applied to a study of inflammation and host response to injury data set derived from the analysis of longitudinal blood microarray profiles of human volunteers treated with intravenous endotoxin compared to placebo. Our approach was able to uncover a cohesive set of differentially expressed genes and novel members exhibiting previously studied differential profiles. This guideline serves as a means to support decisions on new microarray problems.

The rapid development of methods that select over/under expressed genes from microarray experiments have not yet matched the need for tools that identify informational profiles that differentiate between experimental condi tions such as time, treatment and phenotype. Uncertainty arises when methods devoted to identify significantly expressed genes are evaluated: do all microar ray analysis methods yield similar results from the same input dataset? do dif ferent microarray datasets require distinct analysis methods?. We performed a detailed evaluation of several microarray analysis methods, finding that none of these methods alone identifies all observable differential profiles, nor subsumes the results obtained by the other methods. Consequently, we propose a proce dure that, given certain user-defined preferences, generates an optimal suite of statistical methods. These solutions are optimal in the sense that they constitute partial ordered subsets of all possible method-associations bounded by both, the most specific and the most sensitive available solution.