Supervised normalization of microarrays.
Ontology highlight
ABSTRACT:
SUBMITTER: Mecham BH
PROVIDER: S-EPMC2865860 | biostudies-other | 2010 May
REPOSITORIES: biostudies-other
Publications
Supervised normalization of microarrays.
Mecham Brigham H BH Nelson Peter S PS Storey John D JD
Bioinformatics (Oxford, England) 20100331 10
<h4>Motivation</h4>A major challenge in utilizing microarray technologies to measure nucleic acid abundances is 'normalization', the goal of which is to separate biologically meaningful signal from other confounding sources of signal, often due to unavoidable technical factors. It is intuitively clear that true biological signal and confounding factors need to be simultaneously considered when performing normalization. However, the most popular normalization approaches do not utilize what is kno ...[more]
Similar Datasets
Project description:Multiple Arabidopsis thaliana clones from an experimental series of cDNA microarrays are evaluated in order to identify essential sources of noise in the spotting and hybridization process. Theoretical and experimental strategies for an improved quantitative evaluation of cDNA microarrays are proposed and tested on a series of differently diluted control clones. Several sources of noise are identified from the data. Systematic and stochastic fluctuations in the spotting process are reduced by control spots and statistical techniques. The reliability of slide to slide comparison is critically assessed within the statistical framework of pattern matching and classification.
| S-EPMC105386 | biostudies-literature
Project description:BackgroundMiR arrays distinguish themselves from gene expression arrays by their more limited number of probes, and the shorter and less flexible sequence in probe design. Robust data processing and analysis methods tailored to the unique characteristics of miR arrays are greatly needed. Assumptions underlying commonly used normalization methods for gene expression microarrays containing tens of thousands or more probes may not hold for miR microarrays. Findings from previous studies have sometimes been inconclusive or contradictory. Further studies to determine optimal normalization methods for miR microarrays are needed.MethodsWe evaluated many different normalization methods for data generated with a custom-made two channel miR microarray using two data sets that have technical replicates from several different cell lines. The impact of each normalization method was examined on both within miR error variance (between replicate arrays) and between miR variance to determine which normalization methods minimized differences between replicate samples while preserving differences between biologically distinct miRs.ResultsLowess normalization generally did not perform as well as the other methods, and quantile normalization based on an invariant set showed the best performance in many cases unless restricted to a very small invariant set. Global median and global mean methods performed reasonably well in both data sets and have the advantage of computational simplicity.ConclusionsResearchers need to consider carefully which assumptions underlying the different normalization methods appear most reasonable for their experimental setting and possibly consider more than one normalization approach to determine the sensitivity of their results to normalization method used.
| S-EPMC2917409 | biostudies-literature
Project description:Background Array-based comparative genome hybridization (aCGH) is commonly used to determine the genomic content of bacterial strains. In aCGH data, systematic errors are comparable to those occurring in transcriptome data. However, especially for microbes, an additional source of variation exists: differences in hybridization due to gene sequence divergence between the strains hybridized. Current normalization methods do not take this source of variation into consideration. Results We present Supervised Lowess, or S-Lowess, an application of the subset Lowess normalization method that does take difference in genomic content into account. The performance of S-Lowess was assessed on aCGH experiments in which differentially labeled genomic DNA fragments of Lactococcus lactis IL1403 and L. lactis MG1363 strains were hybridized to IL1403 microarrays. Since both genome sequences are known (they have only on average 85 % sequence identity), the success rate in detecting deletions in the MG1363 genome of different aCGH normalization methods can be compared. S-Lowess detects 97% of the deletions, whereas other aCGH normalization methods detect up to only 60% of the deletions. Conclusions S-Lowess removes systematic errors from dual-dye aCGH data in two steps: (i) determination of likely homologous genes (LHG); and (ii) estimation of correction factors for systematic errors from spots of LHG and subset Lowess normalization of the remaining spots using these correction factors. It is implemented in a user-friendly web-tool accessible from http://bioinformatics.biol.rug.nl/websoftware/s-lowess. We demonstrate that it outperforms existing normalization methods and maximizes the number of detectable genomic deletions or duplications from microbial aCGH data. Keywords: comparative genome hybridisation In this study, 4 aCGH comparisons (slides) between L. lactis MG1363 and L. lactis IL1403 were performed (including dye swap with biological replicates; see also supplementary materials). The resulting aCGH slide signals were normalized using the different 'likely homologous gene' (LHG) sets (see above) yielding ratios of signals of labelled DNA of MG1363 over those of IL1403. A maximum of 8 ratios per amplicon (gene) were obtained for the 4 hybridized slides (each with 2 replicate spots per amplicon). Only genes with at least 5 measurements were used in this study. The normalization methods evaluated in this study are: a) no normalization, b) total signal normalization, c) grid-based Lowess (implemented in PreP; f = 0.7) [GarcM-CM--a de la Nava, J et al. 2003. Bioinformatics 19:2328-2329], and d) S-Lowess using different subsets of conserved lactococcal genes (for details see above). Results with the MANOR R package (spatial normalization; standard parameters) [Neuvial P, et al. 2006. BMC Bioinformatics 7:264; Liva S, et al. 2006. Nucleic Acids Res 34:W477-W481] are shown in our supplementary materials.
2008-07-19 | E-GEOD-9039 | biostudies-arrayexpress
Project description:Background Array-based comparative genome hybridization (aCGH) is commonly used to determine the genomic content of bacterial strains. In aCGH data, systematic errors are comparable to those occurring in transcriptome data. However, especially for microbes, an additional source of variation exists: differences in hybridization due to gene sequence divergence between the strains hybridized. Current normalization methods do not take this source of variation into consideration. Results We present Supervised Lowess, or S-Lowess, an application of the subset Lowess normalization method that does take difference in genomic content into account. The performance of S-Lowess was assessed on aCGH experiments in which differentially labeled genomic DNA fragments of Lactococcus lactis IL1403 and L. lactis MG1363 strains were hybridized to IL1403 microarrays. Since both genome sequences are known (they have only on average 85 % sequence identity), the success rate in detecting deletions in the MG1363 genome of different aCGH normalization methods can be compared. S-Lowess detects 97% of the deletions, whereas other aCGH normalization methods detect up to only 60% of the deletions. Conclusions S-Lowess removes systematic errors from dual-dye aCGH data in two steps: (i) determination of likely homologous genes (LHG); and (ii) estimation of correction factors for systematic errors from spots of LHG and subset Lowess normalization of the remaining spots using these correction factors. It is implemented in a user-friendly web-tool accessible from http://bioinformatics.biol.rug.nl/websoftware/s-lowess. We demonstrate that it outperforms existing normalization methods and maximizes the number of detectable genomic deletions or duplications from microbial aCGH data. Keywords: comparative genome hybridisation
2008-07-20 | GSE9039 | GEO
Project description:Microarray data are susceptible to a wide-range of artifacts, many of which occur on physical scales comparable to the spatial dimensions of the array. These artifacts introduce biases that are spatially correlated. The ability of current methodologies to detect and correct such biases is limited.We introduce a new approach for analyzing spatial artifacts, termed 'conditional residual analysis for microarrays' (CRAM). CRAM requires a microarray design that contains technical replicates of representative features and a limited number of negative controls, but is free of the assumptions that constrain existing analytical procedures. The key idea is to extract residuals from sets of matched replicates to generate residual images. The residual images reveal spatial artifacts with single-feature resolution. Surprisingly, spatial artifacts were found to coexist independently as additive and multiplicative errors. Efficient procedures for bias estimation were devised to correct the spatial artifacts on both intensity scales. In a survey of 484 published single-channel datasets, variance fell 4- to 12-fold in 5% of the datasets after bias correction. Thus, inclusion of technical replicates in a microarray design affords benefits far beyond what one might expect with a conventional 'n = 5' averaging, and should be considered when designing any microarray for which randomization is feasible.CRAM is implemented as version 2 of the hoptag software package for R, which is included in the Supplementary information.
| S-EPMC2262854 | biostudies-literature
Project description:MotivationCross-(multi)platform normalization of gene-expression microarray data remains an unresolved issue. Despite the existence of several algorithms, they are either constrained by the need to normalize all samples of all platforms together, compromising scalability and reuse, by adherence to the platforms of a specific provider, or simply by poor performance. In addition, many of the methods presented in the literature have not been specifically tested against multi-platform data and/or other methods applicable in this context. Thus, we set out to develop a normalization algorithm appropriate for gene-expression studies based on multiple, potentially large microarray sets collected along multiple platforms and at different times, applicable in systematic studies aimed at extracting knowledge from the wealth of microarray data available in public repositories; for example, for the extraction of Real-World Data to complement data from Randomized Controlled Trials. Our main focus or criterion for performance was on the capacity of the algorithm to properly separate samples from different biological groups.ResultsWe present CuBlock, an algorithm addressing this objective, together with a strategy to validate cross-platform normalization methods. To validate the algorithm and benchmark it against existing methods, we used two distinct data sets, one specifically generated for testing and standardization purposes and one from an actual experimental study. Using these data sets, we benchmarked CuBlock against ComBat (Johnson et al., 2007), UPC (Piccolo et al., 2013), YuGene (Lê Cao et al., 2014), DBNorm (Meng et al., 2017), Shambhala (Borisov et al., 2019) and a simple log2 transform as reference. We note that many other popular normalization methods are not applicable in this context. CuBlock was the only algorithm in this group that could always and clearly differentiate the underlying biological groups after mixing the data, from up to six different platforms in this study.AvailabilityCuBlock can be downloaded from https://www.mathworks.com/matlabcentral/fileexchange/77882-cublock.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC8388031 | biostudies-literature
Project description:BACKGROUND: Transcriptome data from quantitative PCR (Q-PCR) and DNA microarrays are typically obtained from a fixed amount of RNA collected per sample. Therefore, variations in tissue cellularity and RNA yield across samples in an experimental series compromise accurate determination of the absolute level of each mRNA species per cell in any sample. Since mRNAs are copied from genomic DNA, the simplest way to express mRNA level would be as copy number per template DNA, or more practically, as copy number per cell. RESULTS: Here we report a method (designated the "Percellome" method) for normalizing the expression of mRNA values in biological samples. It provides a "per cell" readout in mRNA copy number and is applicable to both quantitative PCR (Q-PCR) and DNA microarray studies. The genomic DNA content of each sample homogenate was measured from a small aliquot to derive the number of cells in the sample. A cocktail of five external spike RNAs admixed in a dose-graded manner (dose-graded spike cocktail; GSC) was prepared and added to each homogenate in proportion to its DNA content. In this way, the spike mRNAs represented absolute copy numbers per cell in the sample. The signals from the five spike mRNAs were used as a dose-response standard curve for each sample, enabling us to convert all the signals measured to copy numbers per cell in an expression profile-independent manner. A series of samples was measured by Q-PCR and Affymetrix GeneChip microarrays using this Percellome method, and the results showed up to 90 % concordance. CONCLUSION: Percellome data can be compared directly among samples and among different studies, and between different platforms, without further normalization. Therefore, "percellome" normalization can serve as a standard method for exchanging and comparing data across different platforms and among different laboratories.
| S-EPMC1448209 | biostudies-other
Project description:BACKGROUND: Machine learning is a powerful approach for describing and predicting classes in microarray data. Although several comparative studies have investigated the relative performance of various machine learning methods, these often do not account for the fact that performance (e.g. error rate) is a result of a series of analysis steps of which the most important are data normalization, gene selection and machine learning. RESULTS: In this study, we used seven previously published cancer-related microarray data sets to compare the effects on classification performance of five normalization methods, three gene selection methods with 21 different numbers of selected genes and eight machine learning methods. Performance in term of error rate was rigorously estimated by repeatedly employing a double cross validation approach. Since performance varies greatly between data sets, we devised an analysis method that first compares methods within individual data sets and then visualizes the comparisons across data sets. We discovered both well performing individual methods and synergies between different methods. CONCLUSION: Support Vector Machines with a radial basis kernel, linear kernel or polynomial kernel of degree 2 all performed consistently well across data sets. We show that there is a synergistic relationship between these methods and gene selection based on the T-test and the selection of a relatively high number of genes. Also, we find that these methods benefit significantly from using normalized data, although it is hard to draw general conclusions about the relative performance of different normalization procedures.
| S-EPMC3229535 | biostudies-literature
Project description:BACKGROUND: Array-based comparative genome hybridization (aCGH) is commonly used to determine the genomic content of bacterial strains. Since prokaryotes in general have less conserved genome sequences than eukaryotes, sequence divergences between the genes in the genomes used for an aCGH experiment obstruct determination of genome variations (e.g. deletions). Current normalization methods do not take into consideration sequence divergence between target and microarray features and therefore cannot distinguish a difference in signal due to systematic errors in the data or due to sequence divergence. RESULTS: We present supervised Lowess, or S-Lowess, an application of the subset Lowess normalization method. By using a predicted subset of array features with minimal sequence divergence between the analyzed strains for the normalization procedure we remove systematic errors from dual-dye aCGH data in two steps: (1) determination of a subset of conserved genes (i.e. likely conserved genes, LCG); and (2) using the LCG for subset Lowess normalization. Subset Lowess determines the correction factors for systematic errors in the subset of array features and normalizes all array features using these correction factors. The performance of S-Lowess was assessed on aCGH experiments in which differentially labeled genomic DNA fragments of Lactococcus lactis IL1403 and L. lactis MG1363 strains were hybridized to IL1403 DNA microarrays. Since both genomes are sequenced and gene deletions identified, the success rate of different aCGH normalization methods in detecting these deletions in the MG1363 genome were determined. S-Lowess detects 97% of the deletions, whereas other aCGH normalization methods detect up to only 60% of the deletions. CONCLUSION: S-Lowess is implemented in a user-friendly web-tool accessible from http://bioinformatics.biol.rug.nl/websoftware/s-lowess. We demonstrate that it outperforms existing normalization methods and maximizes detection of genomic variation (e.g. deletions) from microbial aCGH data.
| S-EPMC2275246 | biostudies-literature
Project description:Microarray experiments are becoming increasingly common in biomedical research, as is their deposition in publicly accessible repositories, such as Gene Expression Omnibus (GEO). As such, there has been a surge in interest to use this microarray data for meta-analytic approaches, whether to increase sample size for a more powerful analysis of a specific disease (e.g. lung cancer) or to re-examine experiments for reasons different than those examined in the initial, publishing study that generated them. For the average biomedical researcher, there are a number of practical barriers to conducting such meta-analyses such as manually aggregating, filtering and formatting the data. Methods to automatically process large repositories of microarray data into a standardized, directly comparable format will enable easier and more reliable access to microarray data to conduct meta-analyses.We present a straightforward, simple but robust against potential outliers method for automatic quality control and pre-processing of tens of thousands of single-channel microarray data files. GEO GDS files are quality checked by comparing parametric distributions and quantile normalized to enable direct comparison of expression level for subsequent meta-analyses.13,000 human 1-color experiments were processed to create a single gene expression matrix that subsets can be extracted from to conduct meta-analyses. Interestingly, we found that when conducting a global meta-analysis of gene-gene co-expression patterns across all 13,000 experiments to predict gene function, normalization had minimal improvement over using the raw data.Normalization of microarray data appears to be of minimal importance on analyses based on co-expression patterns when the sample size is on the order of thousands microarray datasets. Smaller subsets, however, are more prone to aberrations and artefacts, and effective means of automating normalization procedures not only empowers meta-analytic approaches, but aids in reproducibility by providing a standard way of approaching the problem.Data availability: matrix containing normalized expression of 20,813 genes across 13,000 experiments is available for download at . Source code for GDS files pre-processing is available from the authors upon request.
| S-EPMC3236842 | biostudies-literature