Removing Batch Effects in Analysis of Expression Microarray Data: An Evaluation of Six Batch Adjustment Methods

Abstract

The expression microarray is a frequently used approach to study gene expression on a genome-wide scale. However, the data produced by the thousands of microarray studies published annually are confounded by “batch effects,” the systematic error introduced when samples are processed in multiple batches. Although batch effects can be reduced by careful experimental design, they cannot be eliminated unless the whole study is done in a single batch. A number of programs are now available to adjust microarray data for batch effects prior to analysis. We systematically evaluated six of these programs using multiple measures of precision, accuracy and overall performance. ComBat, an Empirical Bayes method, outperformed the other five programs by most metrics. We also showed that it is essential to standardize expression data at the probe level when testing for correlation of expression profiles, due to a sizeable probe effect in microarray data that can inflate the correlation among replicates and unrelated samples.

Explanation of Variability and Removal of Confounding Factors from Data through Optimal Transport

“…A number of techniques have been developed for removing batch effects; see, for instance, the two reviews in [7,16]. Here we propose a novel approach based on the theory of optimal transport.…”

“…Classical methodologies include ANOVA, ANCOVA, and MANCOVA [20], the many variations of factor analysis [18] (all of these involve linear regression combined with statistical models), stratification [8], detrending of time series, various methodologies for the removal of batch effects [7,16], and a large number of specialized techniques developed within individual fields-notoriously biostatistics, psychology, sociology, econometrics, and the environmental sciences. An alternative methodology involves the statistical elimination of confounding effects by the design of randomized experiments [17], as opposed to the observational studies where covariates are unavoidable.…”

Metabolomic Characterization of Knockout Mutants in Arabidopsis: Development of a Metabolite Profiling Database for Knockout Mutants in Arabidopsis

“…Our experimental design and data processing highlight the importance of reducing nonbiological variations for large-scale comparative metabolomics. The COMBAT algorithm (Johnson et al, 2007) may not be the best choice for adjusting batch effects in metabolomics data sets (Chen et al, 2011). A number of batch adjustment methods are available (De Livera et al, 2012;Lazar et al, 2013), and we are in the process of comparing these methods for their enhancement of the interpretation of metabolomics data sets.…”

Immunosignature system for diagnosis of cancer

“…No sample was used without a technical replicate to estimate reproducibility; therefore, 21 samples were subsequently excluded because their replicate was removed. We then examined the remaining microarrays for signs of microarray batch bias using ComBat (23,24). Because each microarray was printed in batches, it was possible that a printing failure could result in low-quality arrays or arrays that did not match previously manufactured arrays.…”