A Regression-Based Differential Expression Detection Algorithm for Microarray Studies with Ultra-Low Sample Size

Global gene expression analysis using microarrays and, more recently, RNA-seq, has allowed investigators to understand biological processes at a system level. However, the identification of differentially expressed genes in experiments with small sample size, high dimensionality, and high variance remains challenging, limiting the usability of these tens of thousands of publicly available, and possibly many more unpublished, gene expression datasets. We propose a novel variable selection algorithm for ultra-low-n microarray studies using generalized linear model-based variable selection with a penalized binomial regression algorithm called penalized Euclidean distance (PED). Our method uses PED to build a classifier on the experimental data to rank genes by importance. In place of cross-validation, which is required by most similar methods but not reliable for experiments with small sample size, we use a simulation-based approach to additively build a list of differentially expressed genes from the rank-ordered list. Our simulation-based approach maintains a low false discovery rate while maximizing the number of differentially expressed genes identified, a feature critical for downstream pathway analysis. We apply our method to microarray data from an experiment perturbing the Notch signaling pathway in Xenopus laevis embryos. This dataset was chosen because it showed very little differential expression according to limma, a powerful and widely-used method for microarray analysis. Our method was able to detect a significant number of differentially expressed genes in this dataset and suggest future directions for investigation. Our method is easily adaptable for analysis of data from RNA-seq and other global expression experiments with low sample size and high dimensionality.     

Data rotation improves genomotyping efficiency

Unsequenced bacterial strains can be characterized by comparing their genomic DNA to a sequenced reference genome of the same species. This comparative genomic approach, also called genomotyping, is leading to an increased understanding of bacterial evolution and pathogenesis. It is efficiently accomplished by comparative genomic hybridization on custom-designed cDNA microarrays. The microarray experiment results in fluorescence intensities for reference and sample genome for each gene. The log-ratio of these intensities is usually compared to a cut-off, classifying each gene of the sample genome as a candidate for an absent or present gene with respect to the reference genome. Reducing the usually high rate of false positives in the list of candidates for absent genes is decisive for both time and costs of the experiment. We propose a novel method to improve efficiency of genomotyping experiments in this sense, by rotating the normalized intensity data before setting up the list of candidate genes. We analyze simulated genomotyping data and also re-analyze an experimental data set for comparison and illustration. We approximately halve the proportion of false positives in the list of candidate absent genes for the example comparative genomic hybridization experiment as well as for the simulation experiments.     

Evaluation of experimental designs for two-color cDNA microarrays.

The major goal of two-color cDNA microarray experiments is to measure the relative gene expression level (i.e., relative amount of mRNA) of each gene between samples in studies of gene expression. More specifically, given an N-sample experiment, we need all N(N - 1)/2 relative expression levels of all sample pairs of each gene for identification of the differentially expressed genes and for clustering of gene expression patterns. However, the intensities observed from two-color cDNA microarray experiments do not simply represent the relative gene expression level. They are composed of signal (gene expression level), noise, and other factors. In discussions on the experimental design of two-color cDNA microarray experiments, little attention has been given to the fact that different combinations of test and control samples will produce microarray intensities data with varying intrinsic composition of factors. As a consequence, not all experimental designs for two-color cDNA microarray experiments are able to provide all possible relative gene expression levels. This phenomenon has never been addressed. To obtain all possible relative gene expression levels, a novel method for two-color cDNA microarray experimental design evaluation is necessary that will allow the making of an accurate choice. In this study, we propose a model-based approach to illustrate how the factor composition of microarray intensities changed with different experimental designs in two-color cDNA microarray experiments. By analyzing 12 experimental designs (including 5 general forms), we demonstrate that not all experimental designs are able to provide all possible relative gene expression levels due to the differences in factor composition. Our results indicate that whether an experimental design can provide all possible relative expression levels of all sample pairs for each gene should be the first criterion to be considered in an evaluation of experimental designs for two-color cDNA microarray experiments.     

Genomics and genome-wide association studies: an integrative approach to expression QTL mapping

Expression QTL mapping by integrating genome-wide gene expression and genotype data is a promising approach to identifying functional genetic variation, but is hampered by the large number of multiple comparisons inherent in such studies. A novel approach to addressing multiple testing problems in genome-wide family-based association studies is screening candidate markers using heritability or conditional power. We apply these methods in settings in which microarray gene expression data are used as phenotypes, screening for SNPs near the expressed genes. We perform association analyses for phenotypes using a univariate approach. We also perform simulations on trios with large numbers of causal SNPs to determine the optimal number of markers to use in a screen. We demonstrate that our family-based screening approach performs well in the analysis of integrative genomic datasets and that screening using either heritability or conditional power produces similar, though not identical, results.     

Borrowing information across genes and experiments for improved error variance estimation in microarray data analysis

Statistical inference for microarray experiments usually involves the estimation of error variance for each gene. Because the sample size available for each gene is often low, the usual unbiased estimator of the error variance can be unreliable. Shrinkage methods, including empirical Bayes approaches that borrow information across genes to produce more stable estimates, have been developed in recent years. Because the same microarray platform is often used for at least several experiments to study similar biological systems, there is an opportunity to improve variance estimation further by borrowing information not only across genes but also across experiments. We propose a lognormal model for error variances that involves random gene effects and random experiment effects. Based on the model, we develop an empirical Bayes estimator of the error variance for each combination of gene and experiment and call this estimator BAGE because information is Borrowed Across Genes and Experiments. A permutation strategy is used to make inference about the differential expression status of each gene. Simulation studies with data generated from different probability models and real microarray data show that our method outperforms existing approaches.     

A Positive Stable Frailty Model for Clustered Failure Time Data with Covariate‐Dependent Frailty

Summary In this article, we propose a positive stable shared frailty Cox model for clustered failure time data where the frailty distribution varies with cluster‐level covariates. The proposed model accounts for covariate‐dependent intracluster correlation and permits both conditional and marginal inferences. We obtain marginal inference directly from a marginal model, then use a stratified Cox‐type pseudo‐partial likelihood approach to estimate the regression coefficient for the frailty parameter. The proposed estimators are consistent and asymptotically normal and a consistent estimator of the covariance matrix is provided. Simulation studies show that the proposed estimation procedure is appropriate for practical use with a realistic number of clusters. Finally, we present an application of the proposed method to kidney transplantation data from the Scientific Registry of Transplant Recipients.     

Identifying common prognostic factors in genomic cancer studies: A novel index for censored outcomes

Background  

With the growing number of public repositories for high-throughput genomic data, it is of great interest to combine the results produced by independent research groups. Such a combination allows the identification of common genomic factors across multiple cancer types and provides new insights into the disease process. In the framework of the proportional hazards model, classical procedures, which consist of ranking genes according to the estimated hazard ratio or the p-value obtained from a test statistic of no association between survival and gene expression level, are not suitable for gene selection across multiple genomic datasets with different sample sizes. We propose a novel index for identifying genes with a common effect across heterogeneous genomic studies designed to remain stable whatever the sample size and which has a straightforward interpretation in terms of the percentage of separability between patients according to their survival times and gene expression measurements.     

A Hierarchical Semiparametric Model for Incorporating Intergene Information for Analysis of Genomic Data

Summary For analysis of genomic data, e.g., microarray data from gene expression profiling experiments, the two‐component mixture model has been widely used in practice to detect differentially expressed genes. However, it naïvely imposes strong exchangeability assumptions across genes and does not make active use of a priori information about intergene relationships that is currently available, e.g., gene annotations through the Gene Ontology (GO) project. We propose a general strategy that first generates a set of covariates that summarizes the intergene information and then extends the two‐component mixture model into a hierarchical semiparametric model utilizing the generated covariates through latent nonparametric regression. Simulations and analysis of real microarray data show that our method can outperform the naïve two‐component mixture model.     

Linear Mixed Model Selection for False Discovery Rate Control in Microarray Data Analysis

Summary In a microarray experiment, one experimental design is used to obtain expression measures for all genes. One popular analysis method involves fitting the same linear mixed model for each gene, obtaining gene‐specific p‐values for tests of interest involving fixed effects, and then choosing a threshold for significance that is intended to control false discovery rate (FDR) at a desired level. When one or more random factors have zero variance components for some genes, the standard practice of fitting the same full linear mixed model for all genes can result in failure to control FDR. We propose a new method that combines results from the fit of full and selected linear mixed models to identify differentially expressed genes and provide FDR control at target levels when the true underlying random effects structure varies across genes.     

An automatic block and spot indexing with k-nearest neighbors graph for microarray image analysis

MOTIVATION: In this paper, we propose a fully automatic block and spot indexing algorithm for microarray image analysis. A microarray is a device which enables a parallel experiment of ten to hundreds of thousands of test genes in order to measure gene expression. Due to this huge size of experimental data, automated image analysis is gaining importance in microarray image processing systems. Currently, most of the automated microarray image processing systems require manual block indexing and, in some cases, spot indexing. If the microarray image is large and contains a lot of noise, it is very troublesome work. In this paper, we show it is possible to locate the addresses of blocks and spots by applying the Nearest Neighbors Graph Model. Also, we propose an analytic model for the feasibility of block addressing. Our analytic model is validated by a large body of experimental results. RESULTS: We demonstrate the features of automatic block detection, automatic spot addressing, and correction of the distortion and skewedness of each microarray image.     

Mixture modeling for genome-wide localization of transcription factors

Chromatin immunoprecipitation followed by DNA microarray analysis (ChIP-chip methodology) is an efficient way of mapping genome-wide protein-DNA interactions. Data from tiling arrays encompass DNA-protein interaction measurements on thousands or millions of short oligonucleotides (probes) tiling a whole chromosome or genome. We propose a new model-based method for analyzing ChIP-chip data. The proposed model is motivated by the widely used two-component multinomial mixture model of de novo motif finding. It utilizes a hierarchical gamma mixture model of binding intensities while incorporating inherent spatial structure of the data. In this model, genomic regions belong to either one of the following two general groups: regions with a local protein-DNA interaction (peak) and regions lacking this interaction. Individual probes within a genomic region are allowed to have different localization rates accommodating different binding affinities. A novel feature of this model is the incorporation of a distribution for the peak size derived from the experimental design and parameters. This leads to the relaxation of the fixed peak size assumption that is commonly employed when computing a test statistic for these types of spatial data. Simulation studies and a real data application demonstrate good operating characteristics of the method including high sensitivity with small sample sizes when compared to available alternative methods.     

Aligning experimental design with bioinformatics analysis to meet discovery research objectives.

The utility of genomic technology and bioinformatic analytical support to provide new and needed insight into the molecular basis of disease, development, and diversity continues to grow as more research model systems and populations are investigated. Yet deriving results that meet a specific set of research objectives requires aligning or coordinating the design of the experiment, the laboratory techniques, and the data analysis. The following paragraphs describe several important interdependent factors that need to be considered to generate high quality data from the microarray platform. These factors include aligning oligonucleotide probe design with the sample labeling strategy if oligonucleotide probes are employed, recognizing that compromises are inherent in different sample procurement methods, normalizing 2-color microarray raw data, and distinguishing the difference between gene clustering and sample clustering. These factors do not represent an exhaustive list of technical variables in microarray-based research, but this list highlights those variables that span both experimental execution and data analysis.     

Modeling Liquid Association

Summary In 2002, Ker–Chau Li introduced the liquid association measure to characterize three‐way interactions between genes, and developed a computationally efficient estimator that can be used to screen gene expression microarray data for such interactions. That study, and others published since then, have established the biological validity of the method, and clearly demonstrated it to be a useful tool for the analysis of genomic data sets. To build on this work, we have sought a parametric family of multivariate distributions with the flexibility to model the full range of trivariate dependencies encompassed by liquid association. Such a model could situate liquid association within a formal inferential theory. In this article, we describe such a family of distributions, a trivariate, conditional normal model having Gaussian univariate marginal distributions, and in fact including the trivariate Gaussian family as a special case. Perhaps the most interesting feature of the distribution is that the parameterization naturally parses the three‐way dependence structure into a number of distinct, interpretable components. One of these components is very closely aligned to liquid association, and is developed as a measure we call modified liquid association. We develop two methods for estimating this quantity, and propose statistical tests for the existence of this type of dependence. We evaluate these inferential methods in a set of simulations and illustrate their use in the analysis of publicly available experimental data.     

A Penalized Likelihood Approach for Bivariate Conditional Normal Models for Dynamic Co‐expression Analysis

Summary Gene co‐expressions have been widely used in the analysis of microarray gene expression data. However, the co‐expression patterns between two genes can be mediated by cellular states, as reflected by expression of other genes, single nucleotide polymorphisms, and activity of protein kinases. In this article, we introduce a bivariate conditional normal model for identifying the variables that can mediate the co‐expression patterns between two genes. Based on this model, we introduce a likelihood ratio (LR) test and a penalized likelihood procedure for identifying the mediators that affect gene co‐expression patterns. We propose an efficient computational algorithm based on iterative reweighted least squares and cyclic coordinate descent and have shown that when the tuning parameter in the penalized likelihood is appropriately selected, such a procedure has the oracle property in selecting the variables. We present simulation results to compare with existing methods and show that the LR‐based approach can perform similarly or better than the existing method of liquid association and the penalized likelihood procedure can be quite effective in selecting the mediators. We apply the proposed method to yeast gene expression data in order to identify the kinases or single nucleotide polymorphisms that mediate the co‐expression patterns between genes.     

Segmentation and Estimation for SNP Microarrays: A Bayesian Multiple Change‐Point Approach

Summary High‐density single‐nucleotide polymorphism (SNP) microarrays provide a useful tool for the detection of copy number variants (CNVs). The analysis of such large amounts of data is complicated, especially with regard to determining where copy numbers change and their corresponding values. In this article, we propose a Bayesian multiple change‐point model (BMCP) for segmentation and estimation of SNP microarray data. Segmentation concerns separating a chromosome into regions of equal copy number differences between the sample of interest and some reference, and involves the detection of locations of copy number difference changes. Estimation concerns determining true copy number for each segment. Our approach not only gives posterior estimates for the parameters of interest, namely locations for copy number difference changes and true copy number estimates, but also useful confidence measures. In addition, our algorithm can segment multiple samples simultaneously, and infer both common and rare CNVs across individuals. Finally, for studies of CNVs in tumors, we incorporate an adjustment factor for signal attenuation due to tumor heterogeneity or normal contamination that can improve copy number estimates.     

Group additive regression models for genomic data analysis

One important problem in genomic research is to identify genomic features such as gene expression data or DNA single nucleotide polymorphisms (SNPs) that are related to clinical phenotypes. Often these genomic data can be naturally divided into biologically meaningful groups such as genes belonging to the same pathways or SNPs within genes. In this paper, we propose group additive regression models and a group gradient descent boosting procedure for identifying groups of genomic features that are related to clinical phenotypes. Our simulation results show that by dividing the variables into appropriate groups, we can obtain better identification of the group features that are related to the phenotypes. In addition, the prediction mean square errors are also smaller than the component-wise boosting procedure. We demonstrate the application of the methods to pathway-based analysis of microarray gene expression data of breast cancer. Results from analysis of a breast cancer microarray gene expression data set indicate that the pathways of metalloendopeptidases (MMPs) and MMP inhibitors, as well as cell proliferation, cell growth, and maintenance are important to breast cancer-specific survival.