Roland Somogyi

Genetic Network Inference: From Co-Expression Clustering to Reverse Engineering

MOTIVATION Advances in molecular biological, analytical and computational technologies are enabling us to systematically investigate the complex molecular processes underlying biological systems. In particular, using high-throughput gene expression assays, we are able to measure the output of the gene regulatory network. We aim here to review datamining and modeling approaches for conceptualizing and unraveling the functional relationships implicit in these datasets. Clustering of co-expression profiles allows us to infer shared regulatory inputs and functional pathways. We discuss various aspects of clustering, ranging from distance measures to clustering algorithms and multiple-cluster memberships. More advanced analysis aims to infer causal connections between genes directly, i.e. who is regulating whom and how. We discuss several approaches to the problem of reverse engineering of genetic networks, from discrete Boolean networks, to continuous linear and non-linear models. We conclude that the combination of predictive modeling with systematic experimental verification will be required to gain a deeper insight into living organisms, therapeutic targeting and bioengineering.

Yellow fever vaccine induces integrated multilineage and polyfunctional immune responses

Correlates of immune-mediated protection to most viral and cancer vaccines are still unknown. This impedes the development of novel vaccines to incurable diseases such as HIV and cancer. In this study, we have used functional genomics and polychromatic flow cytometry to define the signature of the immune response to the yellow fever (YF) vaccine 17D (YF17D) in a cohort of 40 volunteers followed for up to 1 yr after vaccination. We show that immunization with YF17D leads to an integrated immune response that includes several effector arms of innate immunity, including complement, the inflammasome, and interferons, as well as adaptive immunity as shown by an early T cell response followed by a brisk and variable B cell response. Development of these responses is preceded, as demonstrated in three independent vaccination trials and in a novel in vitro system of primary immune responses (modular immune in vitro construct [MIMIC] system), by the coordinated up-regulation of transcripts for specific transcription factors, including STAT1, IRF7, and ETS2, which are upstream of the different effector arms of the immune response. These results clearly show that the immune response to a strong vaccine is preceded by coordinated induction of master transcription factors that lead to the development of a broad, polyfunctional, and persistent immune response that integrates all effector cells of the immune system.

A gene expression profile of Alzheimer's disease

Postmortem analysis of brains of patients with Alzheimers disease (AD) has led to diverse theories about the causes of the pathology, suggesting that this complex disease involves multiple physiological changes. In an effort to better understand the variety and integration of these changes, we generated a gene expression profile for AD brain. Comparing affected and unaffected brain regions in nine controls and six AD cases, we showed that 118 of the 7050 sequences on a broadly representative cDNA microarray were differentially expressed in the amygdala and cingulate cortex, two regions affected early in the disease. The identity of these genes suggests the most prominent upregulated physiological correlates of pathology involve chronic inflammation, cell adhesion, cell proliferation, and protein synthesis (31 upregulated genes). Conversely, downregulated correlates of pathology involve signal transduction, energy metabolism, stress response, synaptic vesicle synthesis and function, calcium binding, and cytoskeleton (87 downregulated genes). The results support several separate theories of the causes of AD pathology, as well as add to the list of genes associated with AD. In addition, approximately 10 genes of unknown function were found to correlate with the pathology.

Transcription-Based Prediction of Response to IFNβ Using Supervised Computational Methods

Changes in cellular functions in response to drug therapy are mediated by specific transcriptional profiles resulting from the induction or repression in the activity of a number of genes, thereby modifying the preexisting gene activity pattern of the drug-targeted cell(s). Recombinant human interferon beta (rIFNβ) is routinely used to control exacerbations in multiple sclerosis patients with only partial success, mainly because of adverse effects and a relatively large proportion of nonresponders. We applied advanced data-mining and predictive modeling tools to a longitudinal 70-gene expression dataset generated by kinetic reverse-transcription PCR from 52 multiple sclerosis patients treated with rIFNβ to discover higher-order predictive patterns associated with treatment outcome and to define the molecular footprint that rIFNβ engraves on peripheral blood mononuclear cells. We identified nine sets of gene triplets whose expression, when tested before the initiation of therapy, can predict the response to interferon beta with up to 86% accuracy. In addition, time-series analysis revealed potential key players involved in a good or poor response to interferon beta. Statistical testing of a random outcome class and tolerance to noise was carried out to establish the robustness of the predictive models. Large-scale kinetic reverse-transcription PCR, coupled with advanced data-mining efforts, can effectively reveal preexisting and drug-induced gene expression signatures associated with therapeutic effects.

Mining the gene expression matrix: inferring gene relationships from large scale gene expression data

In order to infer the logical principles underlying biological development and phenotypic change, it is necessary to determine large-scale temporal gene expression patters. To quote Eric Lander, “The mRNA levels sensitively reflect the state of the cell, perhaps uniquely defining cell types, stages, and responses. To decipher the logic of gene regulation, we should aim to be able to monitor the expression level of all genes simultaneously…” (Lander, 1996). One method for accomplishing this involves the use of reverse transcription polymerase chain reaction (RT-PCR) to assay the expression levels of large numbers of genes in a tissue at different time points during development, with a standard protocol. The relative amounts of mRNA produced at these time points provide a gene expression time series for each gene.

The application of Shannon entropy in the identification of putative drug targets

A major challenge in the field of functional genomics is the development of computational techniques for organizing and interpreting large amounts of gene expression data. These methods will be critical for the discovery of new therapeutic drug targets. Here, we present a simple method for determining the most likely drug target candidates from temporal gene expression patterns assayed with reverse-transcription polymerase chain reaction (RT-PCR) and DNA microarrays.

Gene Expression Microarray Data Analysis for Toxicology Profiling

Abstract: When dealing with thousands of genes, all potentially interesting, it is desirable to rank the genes according to their degree of participation in a physiological process. Therefore, genes with the highest Shannon entropy and ERL can be selected as the best toxicity target candidates, permitting preclinical scientists to focus their research and resources on those genes.

Determination of Temporal Expression Patterns for Multiple Genes in the Rat Carotid Artery Injury Model

Vascular injury induces extensive alteration to the extracellular matrix (ECM). These changes contribute to lesion formation and promote cell migration and proliferation. To elucidate ECM response to arterial injury, we used real-time polymerase chain reaction monitoring to quantitate the expression levels of 81 genes involved in the synthesis and breakdown of ECM as well as receptors and signaling proteins that communicate and respond to ECM molecules. The temporal regulation of gene expression in the carotid was measured at 1, 3, 5, 7, 9, 14, and 28 days postinjury. Among the 68 genes that showed detectable expression by our method, 47 (69%) were significantly induced or repressed over time, confirming the extensive ECM gene response in this model. More ECM-related genes (31) were regulated at day 1 than at any other time point, and the number of regulated genes decreased over time. However, 14 of the genes were still induced or repressed at day 28, indicating that return to preinjury expression patterns did not occur and no new steady state was achieved over 28 days. In spite of the large number of changes in gene expression, only a small number of expression patterns was observed, suggesting that ECM-related genes could potentially be coregulated.

Making sense of gene-expression data

Large-scale gene-expression analysis aims to increase the depth of diagnostic and drug-effect profiling, and to accelerate the discovery of key biological processes for therapeutic targeting. Success hinges on the precision and scope of the measurements, and on the integration of computational tools for data mining, visualization and modeling. The ultimate objective is to discover key molecular connections in a disease process by reverse engineering signaling networks from their activity profiles.

Reverse engineers map the molecular switching yards

Abstract Networks and pathways are usually regulated and controlled by feedback and feed-forward control – whether the directional behaviors are well understood or not. The consequences and practical opportunities of characterizing and elucidating networks offer enormous potential for biology, medicine and biotechnology. The recent paper by Yeung et al. stands as a fine example of practical reverse engineering method.