Construction of predictive promoter models on the example of antibacterial response of human epithelial cells

Identification of HMX1 target genes: a predictive promoter model approach

PURPOSE

A homozygous mutation in the H6 family homeobox 1 (HMX1) gene is responsible for a new oculoauricular defect leading to eye and auricular developmental abnormalities as well as early retinal degeneration (MIM 612109). However, the HMX1 pathway remains poorly understood, and in the first approach to better understand the pathway's function, we sought to identify the target genes.

METHODS

We developed a predictive promoter model (PPM) approach using a comparative transcriptomic analysis in the retina at P15 of a mouse model lacking functional Hmx1 (dmbo mouse) and its respective wild-type. This PPM was based on the hypothesis that HMX1 binding site (HMX1-BS) clusters should be more represented in promoters of HMX1 target genes. The most differentially expressed genes in the microarray experiment that contained HMX1-BS clusters were used to generate the PPM, which was then statistically validated. Finally, we developed two genome-wide target prediction methods: one that focused on conserving PPM features in human and mouse and one that was based on the co-occurrence of HMX1-BS pairs fitting the PPM, in human or in mouse, independently.

RESULTS

The PPM construction revealed that sarcoglycan, gamma (35kDa dystrophin-associated glycoprotein) (Sgcg), teashirt zinc finger homeobox 2 (Tshz2), and solute carrier family 6 (neurotransmitter transporter, glycine) (Slc6a9) genes represented Hmx1 targets in the mouse retina at P15. Moreover, the genome-wide target prediction revealed that mouse genes belonging to the retinal axon guidance pathway were targeted by Hmx1. Expression of these three genes was experimentally validated using a quantitative reverse transcription PCR approach. The inhibitory activity of Hmx1 on Sgcg, as well as protein tyrosine phosphatase, receptor type, O (Ptpro) and Sema3f, two targets identified by the PPM, were validated with luciferase assay.

CONCLUSIONS

Gene expression analysis between wild-type and dmbo mice allowed us to develop a PPM that identified the first target genes of Hmx1.

N,N-Bis-(8-hydroxyquinoline-5-yl methyl)-benzyl substituted amines (HQNBA): peroxisome proliferator-activated receptor (PPAR-γ) agonists with neuroprotective properties

We report on the neuroprotective effects of N,N-bis-(8-hydroxyquinoline-5-yl methyl)-benzyl substituted amines (HQNBA) in a model of oxidative stress-induced nerve cell death using mouse hippocampal-derived HT22 cells. The four derivatives (JLK1472, JLK1486, JLK1522 and JLK1535) protected the HT22 cells from death at concentrations ranging from 0.1 to 1 μM. Their action is partially dependent on their ability to act as PPARγ agonists. These analogues also maintain GSH levels suggesting that they have indirect anti-oxidant effects.

In silico promoters: modelling of cis-regulatory context facilitates target predictio

Elucidation of gene regulatory complexity holds much promise towards aiding therapeutic interventions in medical research. It has become progressively more evident that the characterization of highly conserved regulatory modules within promoters may assist in the elucidation of distinct cis-motif and trans-element regulatory interactions, shared in response to stimulus-evoked pathological changes. With special emphasis on the promoter, accurate analyses of cis-motif architecture combined with integrative in silico modelling might serve as a more refined approach for prediction and study of regulatory targets and major regulators governing transcriptional control. In this review, we have highlighted key examples and recent advances implementing in silico promoter models that could serve as essential contributions for future research in molecular medicine.

Genome wide prediction of HNF4alpha functional binding sites by the use of local and global sequence context

An application of machine learning algorithms enables prediction of the functional context of transcription factor binding sites in the human genome.

We report an application of machine learning algorithms that enables prediction of the functional context of transcription factor binding sites in the human genome. We demonstrate that our method allowed de novo identification of hepatic nuclear factor (HNF)4α binding sites and significantly improved an overall recognition of faithful HNF4α targets. When applied to published findings, an unprecedented high number of false positives were identified. The technique can be applied to any transcription factor.

Integrative content-driven concepts for bioinformatics “beyond the cell”

Bioinformatics has delivered great contributions to genome and genomics research, without which the world-wide success of this and other global ('omics') approaches would not have been possible. More recently, it has developed further towards the analysis of different kinds of networks thus laying the foundation for comprehensive description, analysis and manipulation of whole living systems in modern "systems biology". The next step which is necessary for developing a systems biology that deals with systemic phenomena is to expand the existing and develop new methodologies that are appropriate to characterize intercellular processes and interactions without omitting the causal underlying molecular mechanisms. Modelling the processes on the different levels of complexity involved requires a comprehensive integration of information on gene regulatory events, signal transduction pathways, protein interaction and metabolic networks as well as cellular functions in the respective tissues / organs.

Composite Module Analyst: identification of transcription factor binding site combinations using genetic algorithm

Composite Module Analyst (CMA) is a novel software tool aiming to identify promoter-enhancer models based on the composition of transcription factor (TF) binding sites and their pairs. CMA is closely interconnected with the TRANSFAC database. In particular, CMA uses the positional weight matrix (PWM) library collected in TRANSFAC and therefore provides the possibility to search for a large variety of different TF binding sites. We model the structure of the long gene regulatory regions by a Boolean function that joins several local modules, each consisting of co-localized TF binding sites. Having as an input a set of co-regulated genes, CMA builds the promoter model and optimizes the parameters of the model automatically by applying a genetic-regression algorithm. We use a multicomponent fitness function of the algorithm which includes several statistical criteria in a weighted linear function. We show examples of successful application of CMA to a microarray data on transcription profiling of TNF-alpha stimulated primary human endothelial cells. The CMA web server is freely accessible at http://www.gene-regulation.com/pub/programs/cma/CMA.html. An advanced version of CMA is also a part of the commercial system ExPlaintrade mark (www.biobase.de) designed for causal analysis of gene expression data.

TiProD: the Tissue-specific Promoter Database

TiProD is a database of human promoter sequences for which some functional features are known. It allows a user to query individual promoters and the expression pattern they mediate, gene expression signatures of individual tissues, and to retrieve sets of promoters according to their tissue-specific activity or according to individual Gene Ontology terms the corresponding genes are assigned to. We have defined a measure for tissue-specificity that allows the user to discriminate between ubiquitously and specifically expressed genes. The database is accessible at http://tiprod.cbi.pku.edu.cn:8080/index.html.

Composite Module Analyst: a fitness-based tool for identification of transcription factor binding site combinations

MOTIVATION

Functionally related genes involved in the same molecular-genetic, biochemical or physiological process are often regulated coordinately. Such regulation is provided by precisely organized binding of a multiplicity of special proteins [transcription factors (TFs)] to their target sites (cis-elements) in regulatory regions of genes. Cis-element combinations provide a structural basis for the generation of unique patterns of gene expression.

RESULTS

Here we present a new approach for defining promoter models based on the composition of TF binding sites and their pairs. We utilize a multicomponent fitness function for selection of the promoter model that fits best to the observed gene expression profile. We demonstrate examples of successful application of the fitness function with the help of a genetic algorithm for the analysis of functionally related or co-expressed genes as well as testing on simulated and permutated data.

AVAILABILITY

The CMA program is freely available for non-commercial users. URL http://www.gene-regulation.com/pub/programs.html#CMAnalyst. It is also a part of the commercial system ExPlain (www.biobase.de) designed for causal analysis of gene expression data..