Exploration, normalization, and summaries of high density oligonucleotide array probe level data
TL;DR: There is no obvious downside to using RMA and attaching a standard error (SE) to this quantity using a linear model which removes probe-specific affinities, and the exploratory data analyses of the probe level data motivate a new summary measure that is a robust multi-array average (RMA) of background-adjusted, normalized, and log-transformed PM values.
Abstract: SUMMARY In this paper we report exploratory analyses of high-density oligonucleotide array data from the Affymetrix GeneChip R � system with the objective of improving upon currently used measures of gene expression. Our analyses make use of three data sets: a small experimental study consisting of five MGU74A mouse GeneChip R � arrays, part of the data from an extensive spike-in study conducted by Gene Logic and Wyeth’s Genetics Institute involving 95 HG-U95A human GeneChip R � arrays; and part of a dilution study conducted by Gene Logic involving 75 HG-U95A GeneChip R � arrays. We display some familiar features of the perfect match and mismatch probe ( PM and MM )v alues of these data, and examine the variance–mean relationship with probe-level data from probes believed to be defective, and so delivering noise only. We explain why we need to normalize the arrays to one another using probe level intensities. We then examine the behavior of the PM and MM using spike-in data and assess three commonly used summary measures: Affymetrix’s (i) average difference (AvDiff) and (ii) MAS 5.0 signal, and (iii) the Li and Wong multiplicative model-based expression index (MBEI). The exploratory data analyses of the probe level data motivate a new summary measure that is a robust multiarray average (RMA) of background-adjusted, normalized, and log-transformed PM values. We evaluate the four expression summary measures using the dilution study data, assessing their behavior in terms of bias, variance and (for MBEI and RMA) model fit. Finally, we evaluate the algorithms in terms of their ability to detect known levels of differential expression using the spike-in data. We conclude that there is no obvious downside to using RMA and attaching a standard error (SE) to this quantity using a linear model which removes probe-specific affinities. ∗ To whom correspondence should be addressed
Citations
More filters
TL;DR: Three methods of performing normalization at the probe intensity level are presented: a one number scaling based algorithm and a method that uses a non-linear normalizing relation by comparing the variability and bias of an expression measure and the simplest and quickest complete data method is found to perform favorably.
Abstract: Motivation: When running experiments that involve multiple high density oligonucleotide arrays, it is important to remove sources of variation between arrays of non-biological origin. Normalization is a process for reducing this variation. It is common to see non-linear relations between arrays and the standard normalization provided by Affymetrix does not perform well in these situations. Results: We present three methods of performing normalization at the probe intensity level. These methods are called complete data methods because they make use of data from all arrays in an experiment to form the normalizing relation. These algorithms are compared to two methods that make use of a baseline array: a one number scaling based algorithm and a method that uses a non-linear normalizing relation by comparing the variability and bias of an expression measure. Two publicly available datasets are used to carry out the comparisons. The simplest and quickest complete data method is found to perform favorably. Availabilty: Software implementing all three of the complete data normalization methods is available as part of the R package Affy, which is a part of the Bioconductor project http://www.bioconductor.org. Contact: bolstad@stat.berkeley.edu Supplementary information: Additional figures may be found at http://www.stat.berkeley.edu/∼bolstad/normalize/ index.html
8,324 citations
Cites background or methods from "Exploration, normalization, and sum..."
...The intensity information from the values of each of the probes in a probeset are combined together to get an expression measure, for example, Average Difference (AvgDiff), the Model Based Expression Index (MBEI) of Li and Wong (2001), the MAS 5....
[...]
...0 Statistical algorithm from Affymetrix (2001), and the Robust Multichip Average proposed in Irizarry et al. (2003). The need for normalization arises naturally when dealing with experiments involving multiple arrays. There are two broad characterizations that could be used for the type of variation one might expect to see when comparing arrays: interesting variation and obscuring variation. We would classify biological differences, for example large differences in the expression level of particular genes between a diseased and a normal tissue source, as interesting variation. However, observed expression levels also include variation that is introduced during the process of carrying out the experiment, which could be classified as obscuring variation. Examples of this obscuring variation arise due to differences in sample preparation (for instance labeling differences), production of the arrays and the processing of the arrays (for instance scanner differences). The purpose of normalization is to deal with this obscuring variation. A more complete discussion on the sources of this variation can be found in Harteminket al. (2001)....
[...]
...Irizarry et al. (2003) contains a more complete discussion of the RMA measure, and further papers exploring its properties are under preparation....
[...]
...The intensity information from the values of each of the probes in a probeset are combined together to get an expression measure, for example, Average Difference (AvgDiff), the Model Based Expression Index (MBEI) of Li and Wong (2001), the MAS 5.0 Statistical algorithm from Affymetrix (2001), and the Robust Multichip Average proposed in Irizarry et al....
[...]
...This data has been previously described in Irizarry et al. (2003)....
[...]
TL;DR: This paper proposed parametric and non-parametric empirical Bayes frameworks for adjusting data for batch effects that is robust to outliers in small sample sizes and performs comparable to existing methods for large samples.
Abstract: SUMMARY Non-biological experimental variation or “batch effects” are commonly observed across multiple batches of microarray experiments, often rendering the task of combining data from these batches difficult. The ability to combine microarray data sets is advantageous to researchers to increase statistical power to detect biological phenomena from studies where logistical considerations restrict sample size or in studies that require the sequential hybridization of arrays. In general, it is inappropriate to combine data sets without adjusting for batch effects. Methods have been proposed to filter batch effects from data, but these are often complicated and require large batch sizes (>25) to implement. Because the majority of microarray studies are conducted using much smaller sample sizes, existing methods are not sufficient. We propose parametric and non-parametric empirical Bayes frameworks for adjusting data for batch effects that is robust to outliers in small sample sizes and performs comparable to existing methods for large samples. We illustrate our methods using two example data sets and show that our methods are justifiable, easy to apply, and useful in practice. Software for our method is freely available at: http://biosun1.harvard.edu/complab/batch/.
6,319 citations
TL;DR: A simple and effective method for performing normalization is outlined and dramatically improved results for inferring differential expression in simulated and publicly available data sets are shown.
Abstract: The fine detail provided by sequencing-based transcriptome surveys suggests that RNA-seq is likely to become the platform of choice for interrogating steady state RNA. In order to discover biologically important changes in expression, we show that normalization continues to be an essential step in the analysis. We outline a simple and effective method for performing normalization and show dramatically improved results for inferring differential expression in simulated and publicly available data sets.
6,042 citations
Cites methods from "Exploration, normalization, and sum..."
...The assumptions behind the TMM method are similar to the assumptions commonly made in microarray normalization procedures such as lowess normalization [21] and quantile normalization [22]....
[...]
TL;DR: It is found that the performance of the current version of the default expression measure provided by Affymetrix Microarray Suite can be significantly improved by the use of probe level summaries derived from empirically motivated statistical models.
Abstract: High density oligonucleotide array technology is widely used in many areas of biomedical research for quantitative and highly parallel measurements of gene expression. Affymetrix GeneChip arrays are the most popular. In this technology each gene is typically represented by a set of 11–20 pairs of probes. In order to obtain expression measures it is necessary to summarize the probe level data. Using two extensive spike-in studies and a dilution study, we developed a set of tools for assessing the effectiveness of expression measures. We found that the performance of the current version of the default expression measure provided by Affymetrix Microarray Suite can be significantly improved by the use of probe level summaries derived from empirically motivated statistical models. In particular, improvements in the ability to detect differentially expressed genes are demonstrated.
5,119 citations
Cites background or methods or result from "Exploration, normalization, and sum..."
...Recent results (4,10) suggest that subtracting MM as a way of correcting for non-speci®c binding is not always appropriate....
[...]
...In the log2 scale, the between-array standard deviation (SD) is in general ®ve times smaller than the within-probe set SD (4,7)....
[...]
...nsf/), two sources of cRNA, human liver tissue and a central nervous system cell line (CNS), were hybridized to human arrays (HG-U95A) in a range of dilutions and proportions (4)....
[...]
...However, the equal variance assumption does not hold for GeneChip probe level data, since probes with larger mean intensities have larger variances (4)....
[...]
...The normalization and background correction procedures used are reported elsewhere (4,9)....
[...]
TL;DR: The affy package is an R package of functions and classes for the analysis of oligonucleotide arrays manufactured by Affymetrix that provides the user with extreme flexibility when carrying out an analysis and make it possible to access and manipulate probe intensity data.
Abstract: Motivation: The processing of the Affymetrix GeneChip data has been a recent focus for data analysts. Alternatives to the original procedure have been proposed and some of these new methods are widely used.
Results: The affy package is an R package of functions and classes for the analysis of oligonucleotide arrays manufactured by Affymetrix. The package is currently in its second release, affy provides the user with extreme flexibility when carrying out an analysis and make it possible to access and manipulate probe intensity data. In this paper, we present the main classes and functions in the package and demonstrate how they can be used to process probe-level data. We also demonstrate the importance of probe-level analysis when using the Affymetrix GeneChip platform.
4,822 citations
References
More filters
TL;DR: Three methods of performing normalization at the probe intensity level are presented: a one number scaling based algorithm and a method that uses a non-linear normalizing relation by comparing the variability and bias of an expression measure and the simplest and quickest complete data method is found to perform favorably.
Abstract: Motivation: When running experiments that involve multiple high density oligonucleotide arrays, it is important to remove sources of variation between arrays of non-biological origin. Normalization is a process for reducing this variation. It is common to see non-linear relations between arrays and the standard normalization provided by Affymetrix does not perform well in these situations. Results: We present three methods of performing normalization at the probe intensity level. These methods are called complete data methods because they make use of data from all arrays in an experiment to form the normalizing relation. These algorithms are compared to two methods that make use of a baseline array: a one number scaling based algorithm and a method that uses a non-linear normalizing relation by comparing the variability and bias of an expression measure. Two publicly available datasets are used to carry out the comparisons. The simplest and quickest complete data method is found to perform favorably. Availabilty: Software implementing all three of the complete data normalization methods is available as part of the R package Affy, which is a part of the Bioconductor project http://www.bioconductor.org. Contact: bolstad@stat.berkeley.edu Supplementary information: Additional figures may be found at http://www.stat.berkeley.edu/∼bolstad/normalize/ index.html
8,324 citations
TL;DR: In this article, the authors proposed a method for monitoring the expression levels of a multiplicity of genes by hybridizing a nucleic acid sample to a high density array of oligonucleotide probes and quantifying the hybridized nucleic acids in the array.
Abstract: This invention provides methods of monitoring the expression levels of a multiplicity of genes. The methods involve hybridizing a nucleic acid sample to a high density array of oligonucleotide probes where the high density array contains oligonucleotide probes complementary to subsequences of target nucleic acids in the nucleic acid sample. In one embodiment, the method involves providing a pool of target nucleic acids comprising RNA transcripts of one or more target genes, or nucleic acids derived from the RNA transcripts, hybridizing said pool of nucleic acids to an array of oligonucleotide probes immobilized on surface, where the array comprising more than 100 different oligonucleotides and each different oligonucleotide is localized in a predetermined region of the surface, the density of the different oligonucleotides is greater than about 60 different oligonucleotides per 1 cm2, and the oligonucleotide probes are complementary to the RNA transcripts or nucleic acids derived from the RNA transcripts; and quantifying the hybridized nucleic acids in the array.
4,382 citations
TL;DR: A statistical model is proposed for the probe-level data, and model-based estimates for gene expression indexes are developed, which help to identify and handle cross-hybridizing probes and contaminating array regions.
Abstract: Recent advances in cDNA and oligonucleotide DNA arrays have made it possible to measure the abundance of mRNA transcripts for many genes simultaneously. The analysis of such experiments is nontrivial because of large data size and many levels of variation introduced at different stages of the experiments. The analysis is further complicated by the large differences that may exist among different probes used to interrogate the same gene. However, an attractive feature of high-density oligonucleotide arrays such as those produced by photolithography and inkjet technology is the standardization of chip manufacturing and hybridization process. As a result, probe-specific biases, although significant, are highly reproducible and predictable, and their adverse effect can be reduced by proper modeling and analysis methods. Here, we propose a statistical model for the probe-level data, and develop model-based estimates for gene expression indexes. We also present model-based methods for identifying and handling cross-hybridizing probes and contaminating array regions. Applications of these results will be presented elsewhere.
3,343 citations
Journal Article•
TL;DR: Differentially expressed genes are identified based on adjusted p-values for a multiple testing procedure which strongly controls the family-wise Type I error rate and takes into account the dependence structure between the gene expression levels.
Abstract: DNA microarrays are a new and promising biotechnology which allows the monitoring of expression levels in cells for thousands of genes simultaneously. The present paper describes statistical methods for the identification of differentially expressed genes in replicated cDNA microarray experiments. Although it is not the main focus of the paper, new methods for the important pre-processing steps of image analysis and normalization are proposed. Given suitably normalized data, the biological question of differential expression is restated as a problem in multiple hypothesis testing: the simultaneous test for each gene of the null hypothesis of no association between the expression levels and responses or covariates of interest. Differentially expressed genes are identified based on adjusted p-values for a multiple testing procedure which strongly controls the family-wise Type I error rate and takes into account the dependence structure between the gene expression levels. No specific parametric form is assumed for the distribution of the test statistics and a permutation procedure is used to estimate adjusted p-values. Several data displays are suggested for the visual identification of differentially expressed genes and of important features of these genes. The above methods are applied to microarray data from a study of gene expression in the livers of mice with very low HDL cholesterol levels. The genes identified using data from multiple slides are compared to those identified by recently published single-slide methods.
1,514 citations
TL;DR: A modified amplification protocol is described that minimizes the generation of template-independent product and can therefore generate the desired microgram quantities of message-derived material from 100 ng of total RNA.
Abstract: Effective transcript profiling in animal systems requires isolation of homogenous tissue or cells followed by faithful mRNA amplification. Linear amplification based on cDNA synthesis and in vitro transcription is reported to maintain representation of mRNA levels, however, quantitative data demonstrating this as well as a description of inherent limitations is lacking. We show that published protocols produce a template-independent product in addition to amplifying real target mRNA thus reducing the specific activity of the final product. We describe a modified amplification protocol that minimizes the generation of template-independent product and can therefore generate the desired microgram quantities of message-derived material from 100 ng of total RNA. Application of a second, nested round of cDNA synthesis and in vitro transcription reduces the required starting material to 2 ng of total RNA. Quantitative analysis of these products on Caenorhabditis elegans Affymetrix GeneChips shows that this amplification does not reduce overall sensitivity and has only minor effects on fidelity.
529 citations