Lee Bennett

Assessing Gene Significance from cDNA Microarray Expression Data via Mixed Models

The determination of a list of differentially expressed genes is a basic objective in many cDNA microarray experiments. We present a statistical approach that allows direct control over the percentage of false positives in such a list and, under certain reasonable assumptions, improves on existing methods with respect to the percentage of false negatives. The method accommodates a wide variety of experimental designs and can simultaneously assess significant differences between multiple types of biological samples. Two interconnected mixed linear models are central to the method and provide a flexible means to properly account for variability both across and within genes. The mixed model also provides a convenient framework for evaluating the statistical power of any particular experimental design and thus enables a researcher to a priori select an appropriate number of replicates. We also suggest some basic graphics for visualizing lists of significant genes. Analyses of published experiments studying human cancer and yeast cells illustrate the results.

Genomic interrogation of mechanism(s) underlying cellular responses to toxicants

Assessment of the impact of xenobiotic exposure on human health and disease progression is complex. Knowledge of mode(s) of action, including mechanism(s) contributing to toxicity and disease progression, is valuable for evaluating compounds. Toxicogenomics, the subdiscipline which merges genomics with toxicology, holds the promise to contributing significantly toward the goal of elucidating mechanism(s) by studying genome-wide effects of xenobiotics. Global gene expression profiling, revolutionized by microarray technology and a crucial aspect of a toxicogenomic study, allows measuring transcriptional modulation of thousands of genes following exposure to a xenobiotic. We use our results from previous studies on compounds representing two different classes of xenobiotics (barbiturate and peroxisome proliferator) to discuss the application of computational approaches for analyzing microarray data to elucidate mechanism(s) underlying cellular responses to toxicants. In particular, our laboratory demonstrated that chemical-specific patterns of gene expression can be revealed using cDNA microarrays. Transcript profiling provides discrimination between classes of toxicants, as well as, genome-wide insight into mechanism(s) of toxicity and disease progression. Ultimately, the expectation is that novel approaches for predicting xenobiotic toxicity in humans will emerge from such information.

ATM-Dependent and -Independent Gene Expression Changes in Response to Oxidative Stress, Gamma Irradiation, and UV Irradiation

Abstract Heinloth, A. N., Shackelford, R. E., Innes, C. L., Bennett, L., Li, L., Amin, R. P., Sieber, S. O., Flores, K. G., Bushel, P. R. and Paules, R. S. ATM-Dependent and -Independent Gene Expression Changes in Response to Oxidative Stress, Gamma Irradiation, and UV Irradiation. Radiat. Res. 160, 273–290 (2003). Ataxia telangiectasia (AT) is an autosomal recessive disorder characterized by progressive cerebellar degeneration, immunodeficiencies, telangiectasias, sensitivity to ionizing radiation, and high predisposition for malignancies. The ataxia telangiectasia mutated (ATM) gene encodes a protein (ATM) with serine/threonine kinase activity. DNA-double strand breaks are known to increase its kinase activity. While cells from individuals with AT are attenuated in their G1-, S- and G2-phase cell cycle checkpoint functions in response to γ irradiation and oxidative stress, their response to UV irradiation appears to be equivalent to that of wild-type cells. In this study, we investigated changes in gene expression in response to γ irradiation, oxidative stress, and UV irradiation, focusing on the dependence on ATM. Doses for all three treatments were selected that resulted in roughly an equivalent induction of a G1 checkpoint response and inhibition of progression through S phase. To investigate gene expression changes, logarithmically growing wild-type and AT dermal diploid fibroblasts were exposed to either γ radiation (5 Gy), oxidative stress (75 µM t-butyl-hydroperoxide), or UV radiation (7.5 J/m2), and RNA was harvested 6 h after treatment. Gene expression analysis was performed using the NIEHS Human ToxChip 2.0 with approximately 1900 cDNA clones representing known genes and ESTs. All three treatments resulted in distinct patterns of gene expression changes, as shown previously. ATM-dependent and ATM-independent components were detected within these patterns, as were novel indications of involvement of ATM in regulation of transcription factors such as SP1, AP1 and MTF1.

Identification of Distinct and Common Gene Expression Changes After Oxidative Stress and Gamma and Ultraviolet Radiation

The human genome is exposed to many different kinds of DNA‐damaging agents. While most damage is detected and repaired through complex damage recognition and repair machineries, some damage has the potential to escape these mechanisms. Unrepaired DNA damage can give rise to alterations and mutations in the genome in an individual cell, which can result in malignant transformation, especially when critical genes are deregulated. In this study, we investigated gene expression changes in response to oxidative stress, gamma (γ) radiation, and ultraviolet (UV) radiation and their potential implications in cancer development. Doses were selected for each of the three treatments, based on their ability to cause a similar G1 checkpoint induction and slow down in early S‐phase progression, as reflected by a comparable reduction in cyclin E–associated kinase activity of at least 75% in logarithmically growing human dermal diploid fibroblasts. To investigate gene expression changes, logarithmically growing dermal diploid fibroblasts were exposed to either γ radiation (5 Gy), oxidative stress (75 μM of tert‐butyl hydroperoxide (t‐butyl‐OOH)), or UV radiation (UVC) (7.5 J/m2) and RNA was harvested 6 h after treatment. Gene expression was analyzed using the NIEHS Human ToxChip 2.0 with approximately 1901 cDNA clones representing known genes and expressed sequence tags (ESTs). We were able to identify common and distinct responses in dermal diploid fibroblasts to the three different stimuli used. Within our analysis, gene expression profiles in response to γ radiation and oxidative stress appeared to be more similar than profiles expressed after UV radiation. Interestingly, equivalent cyclin E–associated kinase activity reduction with all the three treatments was associated with greater transcriptional changes after UV radiation than after γ radiation and oxidative stress. While samples treated with UV radiation displayed modulations of their mitogen activated protein kinase (MAPK) pathway, γ radiation had its major influence on cell‐cycle progression in S‐phase and mitosis. In addition, cell cultures from different individuals displayed significant differences in their gene expression responses to DNA damage. Published 2003 Wiley‐Liss, Inc.

MAPS: a microarray project system for gene expression experiment information and data validation

SUMMARY MAPS is a MicroArray Project System for management and interpretation of microarray gene expression experiment information and data. Microarray project information is organized to track experiments and results that are: (1) validated by performing analysis on stored replicate gene expression data; and (2) queried according to the biological classifications of genes deposited on microarray chips.

Computational selection of distinct class- and subclass-specific gene expression signatures

In this investigation we used statistical methods to select genes with expression profiles that partition classes and subclasses of biological samples. Gene expression data corresponding to liver samples from rats treated for 24 h with an enzyme inducer (phenobarbital) or a peroxisome proliferator (clofibrate, gemfibrozil or Wyeth 14,643) were subjected to a modified Z-score test to identify gene outliers and a binomial distribution to reduce the probability of detecting genes as differentially expressed by chance. Hierarchical clustering of 238 statistically valid differentially expressed genes partitioned class-specific gene expression signatures into groups that clustered samples exposed to the enzyme inducer or to peroxisome proliferators. Using analysis of variance (ANOVA) and linear discriminant analysis methods we identified single genes as well as coupled gene expression profiles that separated the phenobarbital from the peroxisome proliferator treated samples and discerned the fibrate (gemfibrozil and clofibrate) subclass of peroxisome proliferators. A comparison of genes ranked by ANOVA with genes assessed as significant by mixedlinear models analysis [J. Comput. Biol. 8 (2001) 625] or ranked by information gain revealed good congruence with the top 10 genes from each statistical method in the contrast between phenobarbital and peroxisome proliferators expression profiles. We propose building upon a classification regimen comprised of analysis of replicate data, outlier diagnostics and gene selection procedures to utilize cDNA microarray data to categorize subclasses of samples exposed to pharmacologic agents.

Gene expression pattern recognition algorithm inferences to classify samples exposed to chemical agents

We present an analysis of pattern recognition procedures used to predict the classes of samples exposed to pharmacologic agents by comparing gene expression patterns from samples treated with two classes of compounds. Rat liver mRNA samples following exposure for 24 hours with phenobarbital or peroxisome proliferators were analyzed using a 1700 rat cDNA microarray platform. Sets of genes that were consistently differentially expressed in the rat liver samples following treatment were stored in the MicroArray Project System (MAPS) database. MAPS identified 238 genes in common that possessed a low probability (P < 0.01) of being randomly detected as differentially expressed at the 95% confidence level. Hierarchical cluster analysis on the 238 genes clustered specific gene expression profiles that separated samples based on exposure to a particular class of compound.

Gene expression profiling of normal breast epithelial cells following treatment with insulin-like growth factor I

Gene expression profiling of normal breast epithelial cells following treatment with insulin-like growth factor I