Genomic dissection for characterization of cancerous oral epithelium tissues using transcription profiling.

Genome-wide and high-throughput functional genomic tools offer the potential of identifying disease-associated genes and dissecting disease regulatory patterns. There is a need for a set of systematic bioinformatic tools that handles efficiently a large number of variables for extracting biological meaning from experimental outputs. We present well-characterized statistical tools to discover genes that are differentially expressed between malignant oral epithelial and normal tissues in microarray experiments and to construct a robust classifier using the identified discriminatory genes. Those tools include Wilks' lambda score, error rate estimated from leave-one out cross-validation (LOOCV) and Fisher Discriminant Analysis (FDA). High Density DNA microarrays and Real Time Quantitative PCR were employed for the generation and validation of the transcription profile of the oral cancer and normal samples. We identified 45 genes that are strongly correlated with malignancy. Of the 45 genes identified, six have been previously implicated in the disease, and two are uncharacterized clones.

Linking physiology and transcriptional profiles by quantitative predictive models.

A methodology for the construction of quantitative, predictive models of physiology from transcriptional profiles is presented. The method utilizes partial least squares (PLS) regression properly modified to allow gene pre-selection based on their signal-to-noise ratio (SNR). The final set of genes is obtained from a consensus ranking of genes across several thousand trials, each carried out with a different set of training samples. The method was tested with transcriptional data from a large-scale microarray study profiling the effects of high-fat diet on the diet-induced obese mouse model C57BL/6J, and the obese-resistant A/J mouse model. Quantitative predictive models were constructed for the age of the C57BL/6J mice and the A/J mice, and for the insulin and leptin levels of the C57Bl/6J mice based on transcriptional data of liver obtained over a 12-week period. Similarly, models for the growth rate of yeast mutants, and the age of Drosophila samples were developed from literature data. Specifically, it is demonstrated that highly predictive models can be constructed with current levels of precision in DNA microarray measurements provided the variation in the physiological measurements is controlled. Genes identified by this method are important for their ability to collectively predict phenotype. The method can be expanded to include various types of physiological or cellular data, thus providing an integrative framework for the construction of predictive models.

Predicting human skin absorption of chemicals: development of a novel quantitative structure activity relationship.

The objective of this study was to construct and validate a quantitative structure-activity relationship model for skin absorption. Such models are valuable tools for screening and prioritization in safety and efficacy evaluation, and risk assessment of drugs and chemicals. A database of 340 chemicals with percutaneous absorption was assembled. Two models were derived from the training set consisting 306 chemicals (90/10 random split). In addition to the experimental K(ow) values, over 300 2D and 3D atomic and molecular descriptors were analyzed using MDL's QsarIS computer program. Subsequently, the models were validated using both internal (leave-one-out) and external validation (test set) procedures. Using the stepwise regression analysis, three molecular descriptors were determined to have significant statistical correlation with K(p) (R2 = 0.8225): logK(ow), X0 (quantification of both molecular size and the degree of skeletal branching), and SsssCH (count of aromatic carbon groups). In conclusion, two models to estimate skin absorption were developed. When compared to other skin absorption QSAR models in the literature, our model incorporated more chemicals and explored a large number of descriptors. Additionally, our models are reasonably predictive and have met both internal and external statistical validations.

Linking hepatic transcriptional changes to high-fat diet induced physiology for diabetes-prone and obese-resistant mice.

Insulin resistance is characterized by high insulin levels and decreased responsiveness of tissues to the clearance of glucose from the bloodstream. This study maintained the diabetes-prone C57BL/6J and obese-resistant A/J mice strains on a high-fat diet for twelve weeks to transcriptionally profile the liver for changes caused by high fat diet. In the eighth week of the experiment, the C57BL/6J mice began exhibiting signs of insulin resistance, while the A/J mice did not show any such indications during the course of the experiment. A regression model of partial least squares between serum insulin measurements and the liver gene expression profile for the C57BL/6J mice on a high-fat diet was constructed in an effort to quantitatively link the physiological measurement with the gene expressions. A series of discriminating genes between high fat and chow fed mice was generated for both the C57BL/6J and A/J strains. These discriminatory genes contain information about the mechanisms responsible for the development of insulin resistance, and the compensation for a high fat diet, respectively. The results identified several genes involved in the development of insulin resistance and serve as a framework for other studies involving other organs affected by this systemic disease.

Genome-wide dynamic transcriptional profiling of the light-to-dark transition in Synechocystis sp. strain PCC 6803.

We report the results of whole-genome transcriptional profiling of the light-to-dark transition with the model photosynthetic prokaryote Synechocystis sp. strain PCC 6803 (Synechocystis). Experiments were conducted by growing Synechocystis cultures to mid-exponential phase and then exposing them to two cycles of light/dark conditions, during which RNA samples were obtained. These samples were probed with a full-genome DNA microarray (3,169 genes, 20 samples) as well as a partial-genome microarray (88 genes, 29 samples). We concluded that (i) 30-min sampling intervals accurately captured transcriptional dynamics throughout the light/dark transition, (ii) 25% of the Synechocystis genes (783 genes) responded positively to the presence of light, and (iii) the response dynamics varied greatly for individual genes, with a delay of up to 120 to 150 min for some genes. Four classes of genes were identified on the basis of their dynamic gene expression profiles: class I (108 genes, 30-min response time), class II (279 genes, 60 to 90 min), class III (258 genes, 120 to 150 min), and class IV (138 genes, 180 min). The dynamics of several transcripts from genes involved in photosynthesis and primary energy generation are discussed. Finally, we applied Fisher discriminant analysis to better visualize the progression of the overall transcriptional program throughout the light/dark transition and to determine those genes most indicative of the lighting conditions during growth.

Mapping physiological states from microarray expression measurements.

MOTIVATION: The increasing use of DNA microarrays to probe cell physiology requires methods for visualizing different expression phenotypes and explicitly connecting individual genes to discriminating expression features. Such methods should be robust and maintain biological interpretability. RESULTS: We propose a method for the mapping of the physiological state of cells and tissues from multidimensional expression data such as those obtained with DNA microarrays. The method uses Fisher discriminant analysis to create a linear projection of gene expression measurements that maximizes the separation of different sample classes. Relative to other typical classification methods, this method provides insights into the discriminating characteristics of expression measurements in terms of the contribution of individual genes to the definition of distinct physiological states. This projection method also facilitates visualization of classification results in a reduced dimensional space. Examples from four different cases demonstrate the ability of the method to produce well-separated groups in the projection space and to identify important genes for defining physiological states. The method can be augmented to also include data from the proteomic and metabolic phenotypes and can be useful in disease diagnosis, drug screening and bioprocessing applications.

Interactive exploration of microarray gene expression patterns in a reduced dimensional space.

The very high dimensional space of gene expression measurements obtained by DNA microarrays impedes the detection of underlying patterns in gene expression data and the identification of discriminatory genes. In this paper we show the use of projection methods such as principal components analysis (PCA) to obtain a direct link between patterns in the genes and patterns in samples. This feature is useful in the initial interactive pattern exploration of gene expression data and data-driven learning of the nature and types of samples. Using oligonucleotide microarray measurements of 40 samples from different normal human tissues, we show that distinct patterns are obtained when the genes are projected on a two-dimensional plane spanned by the loadings of the two major principal components. These patterns define the particular genes associated with a sample class (i.e., tissue). When used separately from the other genes, these class-specific (i.e., tissue-specific) genes in turn define distinct tissue patterns in the projection space spanned by the scores of the two major principal components. In this study, PCA projection facilitated discriminatory gene selection for different tissues and identified tissue-specific gene expression signatures for liver, skeletal muscle, and brain samples. Furthermore, it allowed the classification of nine new samples belonging to these three types using the linear combination of the expression levels of the tissue-specific genes determined from the first set of samples. The application of the technique to other published data sets is also discussed.

Short-term resistance to diet-induced obesity in A/J mice is not associated with regulation of hypothalamic neuropeptides.

To investigate the mechanisms underlying long-term resistance of the A/J mouse strain to diet-induced obesity, we studied, over a period of 4 wk, the expression of uncoupling proteins in brown adipose tissue and the expression of hypothalamic neuropeptides known to regulate energy homeostasis and then used microarray analysis to identify other potentially important hypothalamic peptides. Despite increased caloric intake after 2 days of high-fat feeding, body weights of A/J mice remained stable. On and after 1 wk of high-fat feeding, A/J mice adjusted their food intake to consume the same amount of calories as mice fed a low-fat diet; thus their body weight and insulin, corticosterone, free fatty acid, and glucose levels remained unchanged for 4 wk. We found no changes in hypothalamic expression of several orexigenic and/or anorexigenic neuropeptides known to play an important role in energy homeostasis for the duration of the study. Uncoupling protein-2 mRNA expression in brown adipose tissue, however, was significantly upregulated after 2 days of high-fat feeding and tended to remain elevated for the duration of the 4-wk study. Gene array analysis revealed that several genes are up- or downregulated in response to 2 days and 1 wk of high-fat feeding. Real-time PCR analysis confirmed that expression of the hypothalamic IL-1 pathway (IL-1beta, IL-1 type 1 and 2 receptors, and PPM1b/PP2C-beta, a molecule that has been implicated in the inhibition of transforming growth factor-beta-activated kinase-1-mediated IL-1 action) is altered after 2 days, but not 1 wk, of high-fat feeding. The role of additional molecules discovered by microarray analysis needs to be further explored in the future.