Dataset Information

Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses.

ABSTRACT: Removal of, or adjustment for, batch effects or center differences is generally required when such effects are present in data. In particular, when preparing microarray gene expression data from multiple cohorts, array platforms, or batches for later analyses, batch effects can have confounding effects, inducing spurious differences between study groups. Many methods and tools exist for removing batch effects from data. However, when study groups are not evenly distributed across batches, actual group differences may induce apparent batch differences, in which case batch adjustments may bias, usually deflate, group differences. Some tools therefore have the option of preserving the difference between study groups, e.g. using a two-way ANOVA model to simultaneously estimate both group and batch effects. Unfortunately, this approach may systematically induce incorrect group differences in downstream analyses when groups are distributed between the batches in an unbalanced manner. The scientific community seems to be largely unaware of how this approach may lead to false discoveries.

SUBMITTER: Nygaard V

PROVIDER: S-EPMC4679072 | biostudies-literature | 2016 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses.

Nygaard Vegard V Rødland Einar Andreas EA Hovig Eivind E

Biostatistics (Oxford, England) 20150813 1

Removal of, or adjustment for, batch effects or center differences is generally required when such effects are present in data. In particular, when preparing microarray gene expression data from multiple cohorts, array platforms, or batches for later analyses, batch effects can have confounding effects, inducing spurious differences between study groups. Many methods and tools exist for removing batch effects from data. However, when study groups are not evenly distributed across batches, actual ...[more]

PMID: 26272994

Similar Datasets

Project description:Recombinant human interferon-gamma (Hu-IFN-gamma) produced by Chinese-hamster ovary (CHO) cells was analysed by immunoprecipitation and SDS/PAGE. Up to twelve molecular-mass variants were secreted by this cell line. Three variants were recovered after enzymic removal of all N-linked oligosaccharides or when glycosylation was inhibited by tunicamycin. The presence of three polypeptide forms rather than a single form suggested that proteolytic cleavage had occurred at two sites in both the glycosylated and non-glycosylated forms. Proteolytically cleaved IFN-gamma was more prevalent in cell lysates than in the secreted glycoprotein. In common with naturally produced IFN-gamma, both fully glycosylated IFN-gamma (asparagine residues 28 and 100 occupied) and partially glycosylated product (thought to be substituted at position Asn28) were secreted. This was deduced from the Mr of the glycosylated products and the relative amounts of sialic acid expressed by each variant. In contrast with naturally produced IFN-gamma, non-glycosylated IFN-gamma was also secreted by the transfected CHO cells. When the cells were grown in batch culture in serum-free medium under pH and dissolved-oxygen control, the proportion of non-glycosylated IFN-gamma increased from 3 to 5% after 3 h, to 30% of the total IFN-gamma present after 195 h. This change in the proportion of glycosylated protein produced was not seen when metabolically labelled IFN-gamma was incubated for 96 h with cell-free supernatant from actively growing CHO cells. This implied that an alteration in intracellular glycosylation was occurring rather than a degradation of oligosaccharide side chains after secretion. The decrease in IFN-gamma glycosylation was independent of the glucose concentration in the culture medium, but could be related to specific growth and IFN-gamma production rates, as these declined steadily after 50 h of culture, in line with the increased production of non-glycosylated IFN-gamma.

Dataset Information

Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses.

Publications

Methods that remove batch effects while retaining group differences may lead to exaggerated confidence in downstream analyses.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets