Hinrich W. H. Göhlmann

Model-Based Approaches

The aim of the analysis presented in Chap. 14 is not to detect genes with a significant dose-response relationship but to use parametric dose-response models in order to compare between several compounds or between the characteristics of the dose-response curves for several genes in a single or multiple compound experiment. The basic methodology for parametric dose-response models presented in the chapter was introduced in Chap. 4. In contrast with Chap. 10 in which the classification post-selection procedure was applied to order-restricted ANOVA models, in this chapter, we focus on the case in which several parametric models are fitted to the gene expression data and we discuss model averaging techniques for the estimation of the ED50 parameter.

Classification of Trends in Dose-Response Microarray Experiments Using Information Theory Selection Methods~!2009-02-24~!2009-07-09~!2009-12-23~!

Dose-response microarray experiments consist of monitoring expression levels of thousands of genes with respect to increasing dose of the treatment under investigation. The primary goal of such an experiment is to establish a dose-response relationship, while the secondary goals are to determine the minimum effective dose level and to identify the shape of the dose-response curve. Recently, Lin et al. (1) discussed several testing procedures to test for monotone trend based on isotonic regression of the observed means (2,3). Once a monotone relationship between the gene expression and dose is established, there is a set of R possible monotone models that can be fitted to the data. A selection of the best model from this set allows us to identify both the shape of dose-response curve and the minimum effective dose level. In this paper we focus on classification of dose-response curve shapes using the information theory model selection. In particular, the Order Restricted Information Criterion (ORIC) is discussed for the inference under order restriction. The posterior probability of the model is calculated using information criteria that take into account both the goodness-of-fit and the complexity of the models. The method is applied to a dose-response microarray experiment with 12 arrays (for three samples at each of the four dose levels) with 16,998 genes.

Genomic biomarkers for a binary clinical outcome in early drug development microarray experiments.

In this article, we discuss methods to select three different types of genes (treatment related, response related, or both) and investigate whether they can serve as biomarkers for a binary outcome variable. We consider an extension of the joint model introduced by Lin et al. (2010) and Tilahun et al. (2010) for a continuous response. As the model has certain drawbacks in a binary setting, we also present a way to use classical selection methods to identify subgroups of genes, which are treatment and/or response related. We evaluate their potential to serve as biomarkers by applying DLDA to predict the response level.

Spectral Map Analysis of Microarray Data

Specific aspects of analyzing microarray data are described, which include size and shape of gene expression profiles, taking logarithms, and the biplot graphic for visualizing associations between genes and cells. Three methods of factor analysis are presented that find application to microarray data: principal component analysis, correspondence analysis, and spectral map analysis. It is shown that these three methods differ only in the way the data are preprocessed and that spectral map analysis has advantages over the other two methods. Two graphical devices that are helpful in exploring and interpreting microarray data are also described.

Translation of disease associated gene signatures across tissues

It has recently been shown that disease associated gene signatures can be identified by profiling tissue other than the disease related tissue. In this paper, we investigate gene signatures for Irritable Bowel Syndrome (IBS) using gene expression profiling of both disease related tissue (colon) and surrogate tissue (rectum). Gene specific joint ANOVA models were used to investigate differentially expressed genes between the IBS patients and the healthy controls taken into account both intra and inter tissue dependencies among expression levels of the same gene. Classification algorithms in combination with feature selection methods were used to investigate the predictive power of gene expression levels from the surrogate and the target tissues. We conclude based on the analyses that expression profiles of the colon and the rectum tissue could result in better predictive accuracy if the disease associated genes are known.

Significance Analysis of Dose-Response Microarray Data

The significance analysis of microarrays (SAM) Tusher et al. (Proc Natl Acad Sci 98:5116–5121, 2001) is a widely used testing procedure that estimates the FDR by using permutations under the assumption that all null hypotheses are true. The procedure consists of three components: (1) the adjusted test statistics, (2) an approximation of the distribution of the test statistics based on permutations, and (3) the control of the FDR. In Chap. 8, we discuss the implementation SAM for the setting of dose-response microarray experiments and illustrate how to use the IsoGene package for SAM.

Quality control of Platinum Spike dataset by probe-level mixed models

Benchmark datasets are important for the validation and optimization of the analysis routes. Lately, a new benchmark dataset, Platinum Spike, for the Affymetrix GeneChip experiments has been introduced. We performed a quality check of the Platinum Spike dataset by using probe-level linear mixed models. The results have shown that there are empty probe sets detecting transcripts, spiked in at different concentrations, and, reversely, there are probe sets that do not detect transcripts, spiked in at different concentrations, even though they were designed to do so. We proposed a formal inference procedure for testing the assumption of independence of all technical replicates in the data and concluded that for almost 10% of probe sets arrays cannot be treated independently, which has strong implications for the normalization procedures and testing for the differential expression. The proposed diagnostics procedure is used to facilitate a thorough exploration of gene expression Affymetrix data beyond the preprocessing and differential expression analysis.

Comparison of methods for the selection of genomic biomarkers

In recent years, a lot of attention is placed on the selection and evaluation of biomarkers in microarray experiments. Two sets of biomarkers are of importance, namely therapeutic and prognostic. The therapeutic biomarkers would give us information on the response of the genes to treatment in relation to the response of the clinical outcome to the same treatments, whereas the prognostic biomarkers enable us to predict the clinical outcome irrespective of treatments and other confounding factors. In this paper, we use different methods that allow for both linear and non-linear associations to select prognostic markers for depression, the response.

Computational Methods Aiding Early-Stage Drug Design (Dagstuhl Seminar 13212)

This report documents the program and the outcomes of Dagstuhl Seminar 13212 Computational Methods Aiding Early-Stage Drug Design. The aim of the seminar was to bring scientists working on various aspects of drug discovery, genomic technologies and computational science (e.g., bioinformatics, chemoinformatics, machine learning, and statistics) together to explore how high dimensional data sets created by genomic technologies can be integrated to identify functional manifestations of drug actions on living cells early in the drug discovery process.

Functional Genomic Dose-Response Experiments

In the first part of the book, we discussed different aspects of the analysis of dose-response data such as estimation, inference, and modeling. In the second part of the book, we focus on dose-response microarray experiments. Within the microarray setting, a dose-response experiment has the same structure as described in Part I of the book. The response is the gene expression at a certain dose level. The role of functional genomics, particularly in this setting, is to find indications of both safety and efficacy before the drug is administrated to patients. In Chap. 5, we give an overview about dose-response microarray experiments and their data structure.