Binary state pattern clustering: a digital paradigm for class and biomarker discovery in gene microarray studies of cancer.

Class and biomarker discovery continue to be among the preeminent goals in gene microarray studies of cancer. We have developed a new data mining technique, which we call Binary State Pattern Clustering (BSPC) that is specifically adapted for these purposes, with cancer and other categorical datasets. BSPC is capable of uncovering statistically significant sample subclasses and associated marker genes in a completely unsupervised manner. This is accomplished through the application of a digital paradigm, where the expression level of each potential marker gene is treated as being representative of its discrete functional state. Multiple genes that divide samples into states along the same boundaries form a kind of gene-cluster that has an associated sample-cluster. BSPC is an extremely fast deterministic algorithm that scales well to large datasets. Here we describe results of its application to three publicly available oligonucleotide microarray datasets. Using an alpha-level of 0.05, clusters reproducing many of the known sample classifications were identified along with associated biomarkers. In addition, a number of simulations were conducted using shuffled versions of each of the original datasets, noise-added datasets, as well as completely artificial datasets. The robustness of BSPC was compared to that of three other publicly available clustering methods: ISIS, CTWC and SAMBA. The simulations demonstrate BSPC's substantially greater noise tolerance and confirm the accuracy of our calculations of statistical significance.     

ArrayNorm: comprehensive normalization and analysis of microarray data

SUMMARY: ArrayNorm is a user-friendly, versatile and platform-independent Java application for the visualization, normalization and analysis of two-color microarray data. A variety of normalization options were implemented to remove the systematic and random errors in the data, taking into account the experimental design and the particularities of every slide. In addition, ArrayNorm provides a module for statistical identification of genes with significant changes in expression. AVAILABILITY: The package is freely available for academic and non-profit institutions from http://genome.tugraz.at     

Merged consensus clustering to assess and improve class discovery with microarray data

Background  

One of the most commonly performed tasks when analysing high throughput gene expression data is to use clustering methods to classify the data into groups. There are a large number of methods available to perform clustering, but it is often unclear which method is best suited to the data and how to quantify the quality of the classifications produced.     

Graph-based consensus clustering for class discovery from gene expression data

MOTIVATION: Consensus clustering, also known as cluster ensemble, is one of the important techniques for microarray data analysis, and is particularly useful for class discovery from microarray data. Compared with traditional clustering algorithms, consensus clustering approaches have the ability to integrate multiple partitions from different cluster solutions to improve the robustness, stability, scalability and parallelization of the clustering algorithms. By consensus clustering, one can discover the underlying classes of the samples in gene expression data. RESULTS: In addition to exploring a graph-based consensus clustering (GCC) algorithm to estimate the underlying classes of the samples in microarray data, we also design a new validation index to determine the number of classes in microarray data. To our knowledge, this is the first time in which GCC is applied to class discovery for microarray data. Given a pre specified maximum number of classes (denoted as K(max) in this article), our algorithm can discover the true number of classes for the samples in microarray data according to a new cluster validation index called the Modified Rand Index. Experiments on gene expression data indicate that our new algorithm can (i) outperform most of the existing algorithms, (ii) identify the number of classes correctly in real cancer datasets, and (iii) discover the classes of samples with biological meaning. AVAILABILITY: Matlab source code for the GCC algorithm is available upon request from Zhiwen Yu.     

Comparison of microarray designs for class comparison and class discovery

MOTIVATION: Two-color microarray experiments in which an aliquot derived from a common RNA sample is placed on each array are called reference designs. Traditionally, microarray experiments have used reference designs, but designs without a reference have recently been proposed as alternatives. RESULTS: We develop a statistical model that distinguishes the different levels of variation typically present in cancer data, including biological variation among RNA samples, experimental error and variation attributable to phenotype. Within the context of this model, we examine the reference design and two designs which do not use a reference, the balanced block design and the loop design, focusing particularly on efficiency of estimates and the performance of cluster analysis. We calculate the relative efficiency of designs when there are a fixed number of arrays available, and when there are a fixed number of samples available. Monte Carlo simulation is used to compare the designs when the objective is class discovery based on cluster analysis of the samples. The number of discrepancies between the estimated clusters and the true clusters were significantly smaller for the reference design than for the loop design. The efficiency of the reference design relative to the loop and block designs depends on the relation between inter- and intra-sample variance. These results suggest that if cluster analysis is a major goal of the experiment, then a reference design is preferable. If identification of differentially expressed genes is the main concern, then design selection may involve a consideration of several factors.     

Effect of data normalization on fuzzy clustering of DNA microarray data

Background  

Microarray technology has made it possible to simultaneously measure the expression levels of large numbers of genes in a short time. Gene expression data is information rich; however, extensive data mining is required to identify the patterns that characterize the underlying mechanisms of action. Clustering is an important tool for finding groups of genes with similar expression patterns in microarray data analysis. However, hard clustering methods, which assign each gene exactly to one cluster, are poorly suited to the analysis of microarray datasets because in such datasets the clusters of genes frequently overlap.     

Model selection and efficiency testing for normalization of cDNA microarray data

In this study we present two novel normalization schemes for cDNA microarrays. They are based on iterative local regression and optimization of model parameters by generalized cross-validation. Permutation tests assessing the efficiency of normalization demonstrated that the proposed schemes have an improved ability to remove systematic errors and to reduce variability in microarray data. The analysis also reveals that without parameter optimization local regression is frequently insufficient to remove systematic errors in microarray data.     

FiGS: a filter-based gene selection workbench for microarray data

Background  

The selection of genes that discriminate disease classes from microarray data is widely used for the identification of diagnostic biomarkers. Although various gene selection methods are currently available and some of them have shown excellent performance, no single method can retain the best performance for all types of microarray datasets. It is desirable to use a comparative approach to find the best gene selection result after rigorous test of different methodological strategies for a given microarray dataset.     

An improved procedure for gene selection from microarray experiments using false discovery rate criterion

Background  

A large number of genes usually show differential expressions in a microarray experiment with two types of tissues, and the p-values of a proper statistical test are often used to quantify the significance of these differences. The genes with small p-values are then picked as the genes responsible for the differences in the tissue RNA expressions. One key question is what should be the threshold to consider the p-values small. There is always a trade off between this threshold and the rate of false claims. Recent statistical literature shows that the false discovery rate (FDR) criterion is a powerful and reasonable criterion to pick those genes with differential expression. Moreover, the power of detection can be increased by knowing the number of non-differential expression genes. While this number is unknown in practice, there are methods to estimate it from data. The purpose of this paper is to present a new method of estimating this number and use it for the FDR procedure construction.     

Tumor classification and marker gene prediction by feature selection and fuzzy c-means clustering using microarray data

Background  

Using DNA microarrays, we have developed two novel models for tumor classification and target gene prediction. First, gene expression profiles are summarized by optimally selected Self-Organizing Maps (SOMs), followed by tumor sample classification by Fuzzy C-means clustering. Then, the prediction of marker genes is accomplished by either manual feature selection (visualizing the weighted/mean SOM component plane) or automatic feature selection (by pair-wise Fisher's linear discriminant).     

Evaluation and comparison of gene clustering methods in microarray analysis

MOTIVATION: Microarray technology has been widely applied in biological and clinical studies for simultaneous monitoring of gene expression in thousands of genes. Gene clustering analysis is found useful for discovering groups of correlated genes potentially co-regulated or associated to the disease or conditions under investigation. Many clustering methods including hierarchical clustering, K-means, PAM, SOM, mixture model-based clustering and tight clustering have been widely used in the literature. Yet no comprehensive comparative study has been performed to evaluate the effectiveness of these methods. RESULTS: In this paper, six gene clustering methods are evaluated by simulated data from a hierarchical log-normal model with various degrees of perturbation as well as four real datasets. A weighted Rand index is proposed for measuring similarity of two clustering results with possible scattered genes (i.e. a set of noise genes not being clustered). Performance of the methods in the real data is assessed by a predictive accuracy analysis through verified gene annotations. Our results show that tight clustering and model-based clustering consistently outperform other clustering methods both in simulated and real data while hierarchical clustering and SOM perform among the worst. Our analysis provides deep insight to the complicated gene clustering problem of expression profile and serves as a practical guideline for routine microarray cluster analysis.     

An unsupervised hierarchical dynamic self-organizing approach to cancer class discovery and marker gene identification in microarray data

MOTIVATION: Current Self-Organizing Maps (SOMs) approaches to gene expression pattern clustering require the user to predefine the number of clusters likely to be expected. Hierarchical clustering methods used in this area do not provide unique partitioning of data. We describe an unsupervised dynamic hierarchical self-organizing approach, which suggests an appropriate number of clusters, to perform class discovery and marker gene identification in microarray data. In the process of class discovery, the proposed algorithm identifies corresponding sets of predictor genes that best distinguish one class from other classes. The approach integrates merits of hierarchical clustering with robustness against noise known from self-organizing approaches. RESULTS: The proposed algorithm applied to DNA microarray data sets of two types of cancers has demonstrated its ability to produce the most suitable number of clusters. Further, the corresponding marker genes identified through the unsupervised algorithm also have a strong biological relationship to the specific cancer class. The algorithm tested on leukemia microarray data, which contains three leukemia types, was able to determine three major and one minor cluster. Prediction models built for the four clusters indicate that the prediction strength for the smaller cluster is generally low, therefore labelled as uncertain cluster. Further analysis shows that the uncertain cluster can be subdivided further, and the subdivisions are related to two of the original clusters. Another test performed using colon cancer microarray data has automatically derived two clusters, which is consistent with the number of classes in data (cancerous and normal). AVAILABILITY: JAVA software of dynamic SOM tree algorithm is available upon request for academic use. SUPPLEMENTARY INFORMATION: A comparison of rectangular and hexagonal topologies for GSOM is available from http://www.mame.mu.oz.au/mechatronics/journalinfo/Hsu2003supp.pdf     

Issues in cDNA microarray analysis: quality filtering, channel normalization, models of variations and assessment of gene effects

We consider the problem of comparing the gene expression levels of cells grown under two different conditions using cDNA microarray data. We use a quality index, computed from duplicate spots on the same slide, to filter out outlying spots, poor quality genes and problematical slides. We also perform calibration experiments to show that normalization between fluorescent labels is needed and that the normalization is slide dependent and non-linear. A rank invariant method is suggested to select non-differentially expressed genes and to construct normalization curves in comparative experiments. After normalization the residuals from the calibration data are used to provide prior information on variance components in the analysis of comparative experiments. Based on a hierarchical model that incorporates several levels of variations, a method for assessing the significance of gene effects in comparative experiments is presented. The analysis is demonstrated via two groups of experiments with 125 and 4129 genes, respectively, in Escherichia coli grown in glucose and acetate.     

Comparative analysis of microarray normalization procedures: effects on reverse engineering gene networks

MOTIVATION: An increasingly common application of gene expression profile data is the reverse engineering of cellular networks. However, common procedures to normalize expression profiles generated using the Affymetrix GeneChips technology were originally developed for a rather different purpose, namely the accurate measure of differential gene expression between two or more phenotypes. As a result, current evaluation strategies lack comprehensive metrics to assess the suitability of available normalization procedures for reverse engineering and, in general, for measuring correlation between the expression profiles of a gene pair. RESULTS: We benchmark four commonly used normalization procedures (MAS5, RMA, GCRMA and Li-Wong) in the context of established algorithms for the reverse engineering of protein-protein and protein-DNA interactions. Replicate sample, randomized and human B-cell data sets are used as an input. Surprisingly, our study suggests that MAS5 provides the most faithful cellular network reconstruction. Furthermore, we identify a crucial step in GCRMA responsible for introducing severe artifacts in the data leading to a systematic overestimate of pairwise correlation. This has key implications not only for reverse engineering but also for other methods, such as hierarchical clustering, relying on accurate measurements of pairwise expression profile correlation. We propose an alternative implementation to eliminate such side effect.     

A stable gene selection in microarray data analysis

Background  

Microarray data analysis is notorious for involving a huge number of genes compared to a relatively small number of samples. Gene selection is to detect the most significantly differentially expressed genes under different conditions, and it has been a central research focus. In general, a better gene selection method can improve the performance of classification significantly. One of the difficulties in gene selection is that the numbers of samples under different conditions vary a lot.     

Systematic variation normalization in microarray data to get gene expression comparison unbiased

Normalization removes or minimizes the biases of systematic variation that exists in experimental data sets. This study presents a systematic variation normalization (SVN) procedure for removing systematic variation in two channel microarray gene expression data. Based on an analysis of how systematic variation contributes to variability in microarray data sets, our normalization procedure includes background subtraction determined from the distribution of pixel intensity values from each data acquisition channel and log conversion, linear or non-linear regression, restoration or transformation, and multiarray normalization. In the case when a non-linear regression is required, an empirical polynomial approximation approach is used. Either the high terminated points or their averaged values in the distributions of the pixel intensity values observed in control channels may be used for rescaling multiarray datasets. These pre-processing steps remove systematic variation in the data attributable to variability in microarray slides, assay-batches, the array process, or experimenters. Biologically meaningful comparisons of gene expression patterns between control and test channels or among multiple arrays are therefore unbiased using normalized but not unnormalized datasets.     

DetectiV: visualization,normalization and significance testing for pathogen-detection microarray data

DNA microarrays offer the possibility of testing for the presence of thousands of micro-organisms in a single experiment. However, there is a lack of reliable bioinformatics tools for the analysis of such data. We have developed DetectiV, a package for the statistical software R. DetectiV offers powerful yet simple visualization, normalization and significance testing tools. We show that DetectiV performs better than previously published software on a large, publicly available dataset.     

An a posteriori strategy for enhancing gene discovery in anonymous cDNA microarray experiments

MOTIVATION: Because of the high cost of sequencing, the bulk of gene discovery is performed using anonymous cDNA microarrays. Though the clones on such arrays are easier and cheaper to construct and utilize than unigene and oligonucleotide arrays, they are there in proportion to their corresponding gene expression activity in the tissue being examined. The associated redundancy will be there in any pool of possibly interesting differentially expressed clones identified in a microarray experiment for subsequent sequencing and investigation. An a posteriori sampling strategy is proposed to enhance gene discovery by reducing the impact of the redundancy in the identified pool. RESULTS: The proposed strategy exploits the fact that individual genes that are highly expressed in a tissue are more likely to be present as a number of spots in an anonymous library and, as a direct consequence, are also likely to give higher fluorescence intensity responses when present in a probe in a cDNA microarray experiment. Consequently, spots that respond with low intensities will have a lower redundancy and so should be sequenced in preference to those with the highest intensities. The proposed method, which formalizes how the fluorescence intensity of a spot should be assessed, is validated using actual microarray data, where the sequences of all the clones in the identified pool had been previously determined. For such validations, the concept of a repeat plot is introduced. It is also utilized to visualize and examine different measures for the characterization of fluorescence intensity. In addition, as confirmatory evidence, sequencing from the lowest to the highest intensities in a pool, with all the sequences known, is compared graphically with their random sequencing. The results establish that, in general, the opportunity for gene discovery is enhanced by avoiding the pooling of different biological libraries (because their construction will have involved different hybridization episodes) and concentrating on the clones with lower fluorescence intensities.     

MarC-V: a spreadsheet-based tool for analysis, normalization, and visualization of single cDNA microarray experiments

The comprehensive analysis and visualization of data extracted from cDNA microarrays can be a time-consuming and error-prone process that becomes increasingly tedious with increased number of gene elements on a particular microarray. With the increasingly large number of gene elements on today's microarrays, analysis tools must be developed to meet this challenge. Here, we present MarC-V, a Microsoft Excel spreadsheet tool with Visual Basic macros to automate much of the visualization and calculation involved in the analysis process while providing the familiarity and flexibility of Excel. Automated features of this tool include (i) lower-bound thresholding, (ii) data normalization, (iii) generation of ratio frequency distribution plots, (iv) generation of scatter plots color-coded by expression level, (v) ratio scoring based on intensity measurements, (vi) filtering of data based on expression level or specific gene interests, and (vii) exporting data for subsequent multi-array analysis. MarC-V also has an importing function included for GenePix results (GPR) raw data files.     

Variable selection for model-based high-dimensional clustering and its application to microarray data

Summary .   Variable selection in high-dimensional clustering analysis is an important yet challenging problem. In this article, we propose two methods that simultaneously separate data points into similar clusters and select informative variables that contribute to the clustering. Our methods are in the framework of penalized model-based clustering. Unlike the classical L ₁-norm penalization, the penalty terms that we propose make use of the fact that parameters belonging to one variable should be treated as a natural "group." Numerical results indicate that the two new methods tend to remove noninformative variables more effectively and provide better clustering results than the L ₁-norm approach.