Analyzing the dependence structure of microarray data: a copula–based approach

The main aim of this Ph.D. dissertation is the study of clustering dependent data by means of copula functions with particular emphasis on microarray data. Copula functions are a popular multivariate modeling tool in each field where the multivariate dependence is of great interest and their use in clustering has not been still investigated. The first part of this work contains the review of the literature of clustering methods, copula functions and microarray experiments. The attention focuses on the K–means (Hartigan, 1975; Hartigan and Wong, 1979), the hierarchical (Everitt, 1974) and the model–based (Fraley and Raftery, 1998, 1999, 2000, 2007) clustering techniques because their performance is compared. Then, the probabilistic interpretation of the Sklar’s theorem (Sklar’s, 1959), the estimation methods for copulas like the Inference for Margins (Joe and Xu, 1996) and the Archimedean and Elliptical copula families are presented. In the end, applications of clustering methods and copulas to the genetic and microarray experiments are highlighted. The second part contains the original contribution proposed. A simulation study is performed in order to evaluate the performance of the K–means and the hierarchical bottom–up clustering methods in identifying clusters according to the dependence structure of the data generating process. Different simulations are performed by varying different conditions (e.g., the kind of margins (distinct, overlapping and nested) and the value of the dependence parameter ) and the results are evaluated by means of different measures of performance. In light of the simulation results and of the limits of the two investigated clustering methods, a new clustering algorithm based on copula functions (‘CoClust’ in brief) is proposed. The basic idea, the iterative procedure of the CoClust and the description of the written R functions with their output are given. The CoClust algorithm is tested on simulated data (by varying the number of clusters, the copula models, the dependence parameter value and the degree of overlap of margins) and is compared with the performance of model–based clustering by using different measures of performance, like the percentage of well–identified number of clusters and the not rejection percentage of H0 on . It is shown that the CoClust algorithm allows to overcome all observed limits of the other investigated clustering techniques and is able to identify clusters according to the dependence structure of the data independently of the degree of overlap of margins and the strength of the dependence. The CoClust uses a criterion based on the maximized log–likelihood function of the copula and can virtually account for any possible dependence relationship between observations. Many peculiar characteristics are shown for the CoClust, e.g. its capability of identifying the true number of clusters and the fact that it does not require a starting classification. Finally, the CoClust algorithm is applied to the real microarray data of Hedenfalk et al. (2001) both to the gene expressions observed in three different cancer samples and to the columns (tumor samples) of the whole data matrix.

New approaches to open problems in gene expression microarray data

In the past decade, the advent of efficient genome sequencing tools and high-throughput experimental biotechnology has lead to enormous progress in the life science. Among the most important innovations is the microarray tecnology. It allows to quantify the expression for thousands of genes simultaneously by measurin the hybridization from a tissue of interest to probes on a small glass or plastic slide. The characteristics of these data include a fair amount of random noise, a predictor dimension in the thousand, and a sample noise in the dozens. One of the most exciting areas to which microarray technology has been applied is the challenge of deciphering complex disease such as cancer. In these studies, samples are taken from two or more groups of individuals with heterogeneous phenotypes, pathologies, or clinical outcomes. these samples are hybridized to microarrays in an effort to find a small number of genes which are strongly correlated with the group of individuals. Eventhough today methods to analyse the data are welle developed and close to reach a standard organization (through the effort of preposed International project like Microarray Gene Expression Data -MGED- Society [1]) it is not unfrequant to stumble in a clinician's question that do not have a compelling statistical method that could permit to answer it.The contribution of this dissertation in deciphering disease regards the development of new approaches aiming at handle open problems posed by clinicians in handle specific experimental designs. In Chapter 1 starting from a biological necessary introduction, we revise the microarray tecnologies and all the important steps that involve an experiment from the production of the array, to the quality controls ending with preprocessing steps that will be used into the data analysis in the rest of the dissertation. While in Chapter 2 a critical review of standard analysis methods are provided stressing most of problems that In Chapter 3 is introduced a method to adress the issue of unbalanced design of miacroarray experiments. In microarray experiments, experimental design is a crucial starting-point for obtaining reasonable results. In a two-class problem, an equal or similar number of samples it should be collected between the two classes. However in some cases, e.g. rare pathologies, the approach to be taken is less evident. We propose to address this issue by applying a modified version of SAM [2]. MultiSAM consists in a reiterated application of a SAM analysis, comparing the less populated class (LPC) with 1,000 random samplings of the same size from the more populated class (MPC) A list of the differentially expressed genes is generated for each SAM application. After 1,000 reiterations, each single probe given a "score" ranging from 0 to 1,000 based on its recurrence in the 1,000 lists as differentially expressed. The performance of MultiSAM was compared to the performance of SAM and LIMMA [3] over two simulated data sets via beta and exponential distribution. The results of all three algorithms over low- noise data sets seems acceptable However, on a real unbalanced two-channel data set reagardin Chronic Lymphocitic Leukemia, LIMMA finds no significant probe, SAM finds 23 significantly changed probes but cannot separate the two classes, while MultiSAM finds 122 probes with score >300 and separates the data into two clusters by hierarchical clustering. We also report extra-assay validation in terms of differentially expressed genes Although standard algorithms perform well over low-noise simulated data sets, multi-SAM seems to be the only one able to reveal subtle differences in gene expression profiles on real unbalanced data. In Chapter 4 a method to adress similarities evaluation in a three-class prblem by means of Relevance Vector Machine [4] is described. In fact, looking at microarray data in a prognostic and diagnostic clinical framework, not only differences could have a crucial role. In some cases similarities can give useful and, sometimes even more, important information. The goal, given three classes, could be to establish, with a certain level of confidence, if the third one is similar to the first or the second one. In this work we show that Relevance Vector Machine (RVM) [2] could be a possible solutions to the limitation of standard supervised classification. In fact, RVM offers many advantages compared, for example, with his well-known precursor (Support Vector Machine - SVM [3]). Among these advantages, the estimate of posterior probability of class membership represents a key feature to address the similarity issue. This is a highly important, but often overlooked, option of any practical pattern recognition system. We focused on Tumor-Grade-three-class problem, so we have 67 samples of grade I (G1), 54 samples of grade 3 (G3) and 100 samples of grade 2 (G2). The goal is to find a model able to separate G1 from G3, then evaluate the third class G2 as test-set to obtain the probability for samples of G2 to be member of class G1 or class G3. The analysis showed that breast cancer samples of grade II have a molecular profile more similar to breast cancer samples of grade I. Looking at the literature this result have been guessed, but no measure of significance was gived before.

Mining tissue microarray data to uncover combinations of biomarker expression patterns that improve intermediate staging and grading of clear cell renal cell cancer

PURPOSE: Tumor stage and nuclear grade are the most important prognostic parameters of clear cell renal cell carcinoma (ccRCC). The progression risk of ccRCC remains difficult to predict particularly for tumors with organ-confined stage and intermediate differentiation grade. Elucidating molecular pathways deregulated in ccRCC may point to novel prognostic parameters that facilitate planning of therapeutic approaches. EXPERIMENTAL DESIGN: Using tissue microarrays, expression patterns of 15 different proteins were evaluated in over 800 ccRCC patients to analyze pathways reported to be physiologically controlled by the tumor suppressors von Hippel-Lindau protein and phosphatase and tensin homologue (PTEN). Tumor staging and grading were improved by performing variable selection using Cox regression and a recursive bootstrap elimination scheme. RESULTS: Patients with pT2 and pT3 tumors that were p27 and CAIX positive had a better outcome than those with all remaining marker combinations. A prolonged survival among patients with intermediate grade (grade 2) correlated with both nuclear p27 and cytoplasmic PTEN expression, as well as with inactive, nonphosphorylated ribosomal protein S6. By applying graphical log-linear modeling for over 700 ccRCC for which the molecular parameters were available, only a weak conditional dependence existed between the expression of p27, PTEN, CAIX, and p-S6, suggesting that the dysregulation of several independent pathways are crucial for tumor progression. CONCLUSIONS: The use of recursive bootstrap elimination, as well as graphical log-linear modeling for comprehensive tissue microarray (TMA) data analysis allows the unraveling of complex molecular contexts and may improve predictive evaluations for patients with advanced renal cancer.

Systemic analysis of microarray data sets towards identifying genes regulating root development

Nitrogen and water are essential for plant growth and development. In this study, we designed experiments to produce gene expression data of poplar roots under nitrogen starvation and water deprivation conditions. We found low concentration of nitrogen led first to increased root elongation followed by lateral root proliferation and eventually increased root biomass. To identify genes regulating root growth and development under nitrogen starvation and water deprivation, we designed a series of data analysis procedures, through which, we have successfully identified biologically important genes. Differentially Expressed Genes (DEGs) analysis identified the genes that are differentially expressed under nitrogen starvation or drought. Protein domain enrichment analysis identified enriched themes (in same domains) that are highly interactive during the treatment. Gene Ontology (GO) enrichment analysis allowed us to identify biological process changed during nitrogen starvation. Based on the above analyses, we examined the local Gene Regulatory Network (GRN) and identified a number of transcription factors. After testing, one of them is a high hierarchically ranked transcription factor that affects root growth under nitrogen starvation. It is very tedious and time-consuming to analyze gene expression data. To avoid doing analysis manually, we attempt to automate a computational pipeline that now can be used for identification of DEGs and protein domain analysis in a single run. It is implemented in scripts of Perl and R.

Selecting control genes for RT-QPCR using public microarray data

BACKGROUND: Gene expression analysis has emerged as a major biological research area, with real-time quantitative reverse transcription PCR (RT-QPCR) being one of the most accurate and widely used techniques for expression profiling of selected genes. In order to obtain results that are comparable across assays, a stable normalization strategy is required. In general, the normalization of PCR measurements between different samples uses one to several control genes (e.g. housekeeping genes), from which a baseline reference level is constructed. Thus, the choice of the control genes is of utmost importance, yet there is not a generally accepted standard technique for screening a large number of candidates and identifying the best ones. RESULTS: We propose a novel approach for scoring and ranking candidate genes for their suitability as control genes. Our approach relies on publicly available microarray data and allows the combination of multiple data sets originating from different platforms and/or representing different pathologies. The use of microarray data allows the screening of tens of thousands of genes, producing very comprehensive lists of candidates. We also provide two lists of candidate control genes: one which is breast cancer-specific and one with more general applicability. Two genes from the breast cancer list which had not been previously used as control genes are identified and validated by RT-QPCR. Open source R functions are available at http://www.isrec.isb-sib.ch/~vpopovic/research/ CONCLUSION: We proposed a new method for identifying candidate control genes for RT-QPCR which was able to rank thousands of genes according to some predefined suitability criteria and we applied it to the case of breast cancer. We also empirically showed that translating the results from microarray to PCR platform was achievable.

SURVIVAL PREDICTION FOR BRAIN TUMOR PATIENTS USING GENE EXPRESSION DATA

Brain tumor is one of the most aggressive types of cancer in humans, with an estimated median survival time of 12 months and only 4% of the patients surviving more than 5 years after disease diagnosis. Until recently, brain tumor prognosis has been based only on clinical information such as tumor grade and patient age, but there are reports indicating that molecular profiling of gliomas can reveal subgroups of patients with distinct survival rates. We hypothesize that coupling molecular profiling of brain tumors with clinical information might improve predictions of patient survival time and, consequently, better guide future treatment decisions. In order to evaluate this hypothesis, the general goal of this research is to build models for survival prediction of glioma patients using DNA molecular profiles (U133 Affymetrix gene expression microarrays) along with clinical information. First, a predictive Random Forest model is built for binary outcomes (i.e. short vs. long-term survival) and a small subset of genes whose expression values can be used to predict survival time is selected. Following, a new statistical methodology is developed for predicting time-to-death outcomes using Bayesian ensemble trees. Due to a large heterogeneity observed within prognostic classes obtained by the Random Forest model, prediction can be improved by relating time-to-death with gene expression profile directly. We propose a Bayesian ensemble model for survival prediction which is appropriate for high-dimensional data such as gene expression data. Our approach is based on the ensemble "sum-of-trees" model which is flexible to incorporate additive and interaction effects between genes. We specify a fully Bayesian hierarchical approach and illustrate our methodology for the CPH, Weibull, and AFT survival models. We overcome the lack of conjugacy using a latent variable formulation to model the covariate effects which decreases computation time for model fitting. Also, our proposed models provides a model-free way to select important predictive prognostic markers based on controlling false discovery rates. We compare the performance of our methods with baseline reference survival methods and apply our methodology to an unpublished data set of brain tumor survival times and gene expression data, selecting genes potentially related to the development of the disease under study. A closing discussion compares results obtained by Random Forest and Bayesian ensemble methods under the biological/clinical perspectives and highlights the statistical advantages and disadvantages of the new methodology in the context of DNA microarray data analysis.

THEORETICAL STUDIES ON THE METHODS OF RECONSTRUCTING PHYLOGENETIC TREES FROM DNA SEQUENCE DATA (MOLECULAR EVOLUTION, HOMINOID EVOLUTION, COMPUTER SIMULATION)

(1) A mathematical theory for computing the probabilities of various nucleotide configurations is developed, and the probability of obtaining the correct phylogenetic tree (model tree) from sequence data is evaluated for six phylogenetic tree-making methods (UPGMA, distance Wagner method, transformed distance method, Fitch-Margoliash's method, maximum parsimony method, and compatibility method). The number of nucleotides (m*) necessary to obtain the correct tree with a probability of 95% is estimated with special reference to the human, chimpanzee, and gorilla divergence. m* is at least 4,200, but the availability of outgroup species greatly reduces m* for all methods except UPGMA. m* increases if transitions occur more frequently than transversions as in the case of mitochondrial DNA. (2) A new tree-making method called the neighbor-joining method is proposed. This method is applicable either for distance data or character state data. Computer simulation has shown that the neighbor-joining method is generally better than UPGMA, Farris' method, Li's method, and modified Farris method on recovering the true topology when distance data are used. A related method, the simultaneous partitioning method, is also discussed. (3) The maximum likelihood (ML) method for phylogeny reconstruction under the assumption of both constant and varying evolutionary rates is studied, and a new algorithm for obtaining the ML tree is presented. This method gives a tree similar to that obtained by UPGMA when constant evolutionary rate is assumed, whereas it gives a tree similar to that obtained by the maximum parsimony tree and the neighbor-joining method when varying evolutionary rate is assumed. ^

Integrating sequence information in microarray data analysis by free energy modeling

Microarray technology is a high-throughput method for genotyping and gene expression profiling. Limited sensitivity and specificity are one of the essential problems for this technology. Most of existing methods of microarray data analysis have an apparent limitation for they merely deal with the numerical part of microarray data and have made little use of gene sequence information. Because it's the gene sequences that precisely define the physical objects being measured by a microarray, it is natural to make the gene sequences an essential part of the data analysis. This dissertation focused on the development of free energy models to integrate sequence information in microarray data analysis. The models were used to characterize the mechanism of hybridization on microarrays and enhance sensitivity and specificity of microarray measurements. ^ Cross-hybridization is a major obstacle factor for the sensitivity and specificity of microarray measurements. In this dissertation, we evaluated the scope of cross-hybridization problem on short-oligo microarrays. The results showed that cross hybridization on arrays is mostly caused by oligo fragments with a run of 10 to 16 nucleotides complementary to the probes. Furthermore, a free-energy based model was proposed to quantify the amount of cross-hybridization signal on each probe. This model treats cross-hybridization as an integral effect of the interactions between a probe and various off-target oligo fragments. Using public spike-in datasets, the model showed high accuracy in predicting the cross-hybridization signals on those probes whose intended targets are absent in the sample. ^ Several prospective models were proposed to improve Positional Dependent Nearest-Neighbor (PDNN) model for better quantification of gene expression and cross-hybridization. ^ The problem addressed in this dissertation is fundamental to the microarray technology. We expect that this study will help us to understand the detailed mechanism that determines sensitivity and specificity on the microarrays. Consequently, this research will have a wide impact on how microarrays are designed and how the data are interpreted. ^

Two way clustering of Microarray Data using a Hybrid Approach

The Microarray technique is rather powerful, as it allows to test up thousands of genes at a time, but this produces an overwhelming set of data files containing huge amounts of data, which is quite difficult to pre-process, separate, classify and correlate for interesting conclusions to be extracted. Modern machine learning, data mining and clustering techniques based on information theory, are needed to read and interpret the information contents buried in those large data sets. Independent Component Analysis method can be used to correct the data affected by corruption processes or to filter the uncorrectable one and then clustering methods can group similar genes or classify samples. In this paper a hybrid approach is used to obtain a two way unsupervised clustering for a corrected microarray data.

MethDB—a public database for DNA methylation data

Methylation of cytosine in the 5 position of the pyrimidine ring is a major modification of the DNA in most organisms. In eukaryotes, the distribution and number of 5-methylcytosines (5mC) along the DNA is heritable but can also change with the developmental state of the cell and as a response to modifications of the environment. While DNA methylation probably has a number of functions, scientific interest has recently focused on the gene silencing effect methylation can have in eukaryotic cells. In particular, the discovery of changes in the methylation level during cancer development has increased the interest in this field. In the past, a vast amount of data has been generated with different levels of resolution ranging from 5mC content of total DNA to the methylation status of single nucleotides. We present here a database for DNA methylation data that attempts to unify these results in a common resource. The database is accessible via WWW (http://www.methdb.de). It stores information about the origin of the investigated sample and the experimental procedure, and contains the DNA methylation data. Query masks allow for searching for 5mC content, species, tissue, gene, sex, phenotype, sequence ID and DNA type. The output lists all available information including the relative gene expression level. DNA methylation patterns and methylation profiles are shown both as a graphical representation and as G/A/T/C/5mC-sequences or tables with sequence positions and methylation levels, respectively.

Statistical modeling of large microarray data sets to identify stimulus-response profiles

A statistical modeling approach is proposed for use in searching large microarray data sets for genes that have a transcriptional response to a stimulus. The approach is unrestricted with respect to the timing, magnitude or duration of the response, or the overall abundance of the transcript. The statistical model makes an accommodation for systematic heterogeneity in expression levels. Corresponding data analyses provide gene-specific information, and the approach provides a means for evaluating the statistical significance of such information. To illustrate this strategy we have derived a model to depict the profile expected for a periodically transcribed gene and used it to look for budding yeast transcripts that adhere to this profile. Using objective criteria, this method identifies 81% of the known periodic transcripts and 1,088 genes, which show significant periodicity in at least one of the three data sets analyzed. However, only one-quarter of these genes show significant oscillations in at least two data sets and can be classified as periodic with high confidence. The method provides estimates of the mean activation and deactivation times, induced and basal expression levels, and statistical measures of the precision of these estimates for each periodic transcript.

Recursive partitioning for tumor classification with gene expression microarray data

Precise classification of tumors is critically important for cancer diagnosis and treatment. It is also a scientifically challenging task. Recently, efforts have been made to use gene expression profiles to improve the precision of classification, with limited success. Using a published data set for purposes of comparison, we introduce a methodology based on classification trees and demonstrate that it is significantly more accurate for discriminating among distinct colon cancer tissues than other statistical approaches used heretofore. In addition, competing classification trees are displayed, which suggest that different genes may coregulate colon cancers.