CoChat
Sign Up
Cited by 230Gene expression and cancer classificationGenomics and Phylogenetic StudiesBioinformatics and Genomic Networks

The DOI matches its Crossref record; 32 of the 32 references it cites were found in a registry or an index.

Optimizing the noise versus bias trade-off for Illumina whole genome expression BeadChips

Open access
Shi, Wei, Oshlack, Alicia, Smyth, Gordon K.·2010·Nucleic Acids Research
DOI 10.1093/nar/gkq871PMC PMC3001098
Open this paper in CoChat — free accountOpen PDF

Abstract

Five strategies for pre-processing intensities from Illumina expression BeadChips are assessed from the point of view of precision and bias. The strategies include a popular variance stabilizing transformation and model-based background corrections that either use or ignore the control probes. Four calibration data sets are used to evaluate precision, bias and false discovery rate (FDR). The original algorithms are shown to have operating characteristics that are not easily comparable. Some tend to minimize noise while others minimize bias. Each original algorithm is shown to have an innate intensity offset, by which unlogged intensities are bounded away from zero, and the size of this offset determines its position on the noise-bias spectrum. By adding extra offsets, a continuum of related algorithms with different noise-bias trade-offs is generated, allowing direct comparison of the performance of the strategies on equivalent terms. Adding a positive offset is shown to decrease the FDR of each original algorithm. The potential of each strategy to generate an algorithm with an optimal noise-bias trade-off is explored by finding the offset that minimizes its FDR. The use of control probes as part of the background correction and normalization strategy is shown to achieve the lowest FDR for a given bias.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1093/nar/gkq871

Cites

32 of 32 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1. Operating characteristics of each pre-processing strategy. Shown are boxplots of normalized intensities (a1–a4), sample variances (b1–b4) and absolute fold changes (c1–c4) for each pre-processing strategy applied to each data set. Each boxplot shows the spread of values across microarray probes for a particular strategy applied to a particular data set. The vertical axis has been truncated in some cases to better show the main body of the boxplots. The first and second rows show results for mixture data sets M1 and M2, respectively. Panels c1-c2 show fold changes for comparing pure samples. The third and fourth rows show results for spike-in data sets S1 and S2, respectively. Only results for spike-in probes are shown (note that spike-in probes and array probes were normalized together). Panels a3–a4 show intensities for probes for which the spike-in concentration is actually 0. Panels c3–c4 show fold changes from the comparison between spike-in concentrations 1000 pM and 0.01 pM. In the left column (a1–a4), longer boxes indicate a good range of normalized values. In the middle column (b1–b4), lower boxes indicate higher precision. In the third column (c1–c4), longer and higher boxes indicate a greater range of fold changes.
Figure 1. Operating characteristics of each pre-processing strategy. Shown are boxplots of normalized intensities (a1–a4), sample variances (b1–b4) and absolute fold changes (c1–c4) for each pre-processing strategy applied to each data set. Each boxplot shows the spread of values across microarray probes for a particular strategy applied to a particular data set. The vertical axis has been truncated in some cases to better show the main body of the boxplots. The first and second rows show results for mixture data sets M1 and M2, respectively. Panels c1-c2 show fold changes for comparing pure samples. The third and fourth rows show results for spike-in data sets S1 and S2, respectively. Only results for spike-in probes are shown (note that spike-in probes and array probes were normalized together). Panels a3–a4 show intensities for probes for which the spike-in concentration is actually 0. Panels c3–c4 show fold changes from the comparison between spike-in concentrations 1000 pM and 0.01 pM. In the left column (a1–a4), longer boxes indicate a good range of normalized values. In the middle column (b1–b4), lower boxes indicate higher precision. In the third column (c1–c4), longer and higher boxes indicate a greater range of fold changes.CC BY-NC · PubMed Central
Figure 2. Operating characteristics of pre-processing strategies when total offsets are forced equal. Shown are boxplots of (a) normalized intensities, (b) log variances and (c) absolute fold changes. Results shown are for data set M1, with added offsets of 73, 237 and 256 for the logq, neq and neqc strategies, respectively.
Figure 2. Operating characteristics of pre-processing strategies when total offsets are forced equal. Shown are boxplots of (a) normalized intensities, (b) log variances and (c) absolute fold changes. Results shown are for data set M1, with added offsets of 73, 237 and 256 for the logq, neq and neqc strategies, respectively.CC BY-NC · PubMed Central
Figure 3. Operating characteristics of the logq pre-processing strategy for different added offsets. Shown are boxplots of (a) normalized intensities, (b) variances and (c) absolute fold changes for comparing pure samples. Results shown are for data M1.
Figure 3. Operating characteristics of the logq pre-processing strategy for different added offsets. Shown are boxplots of (a) normalized intensities, (b) variances and (c) absolute fold changes for comparing pure samples. Results shown are for data M1.CC BY-NC · PubMed Central
Figure 4. AUC as a function of added offset for the logq, neq and neqc strategies. Panel (a) is for data set M1, (b) for data set M2.
Figure 4. AUC as a function of added offset for the logq, neq and neqc strategies. Panel (a) is for data set M1, (b) for data set M2.CC BY-NC · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 32 references

Research works printed but not linked: 1

  1. Illumina, 2008, BeadStudio Gene Expression Module User Guide

Cited by

Sorted by citations

See all 230

Related

Metadata

Authors
Shi, W., Oshlack, A., & Smyth, G. K.
Affiliations
The University of Melbourne (AU)
Corresponding author
  • Shi, Wei
  • Smyth, Gordon K.
ORCID
Venue
Nucleic Acids Research
Publisher
Oxford University Press (OUP)
Published
2010-10-06
Volume
38
Issue
22
Pages
e204-e204
ISSN
  • 0305-1048 (ISSN-L)
  • 1362-4962
DOI
10.1093/nar/gkq871
PMC
PMC3001098
OpenAlex
W2118373398
Cited by
225
References
32
Citations by year
2026: 5 · 2025: 4 · 2024: 8 · 2023: 10 · 2022: 13 · 2021: 15 · 2020: 11 · 2019: 20 · 2018: 14 · 2017: 19 · 2016: 28 · 2015: 23 · 2014: 15 · 2013: 20 · 2012: 9
FWCI
5.21
Type
journal article
Language
en
Open access
Licence
Locations
  • Nucleic Acids Research (publishedVersion, cc-by-nc)
  • PubMed Central (publishedVersion, cc-by-nc)
  • PubMed (publishedVersion)
  • Minerva Access (University of Melbourne) (submittedVersion, cc-by-nc)
  • HighWire Press Open Archive (submittedVersion)
Funders
  • Medical Research Council
  • National Health and Medical Research Council (490037)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
Keywords
variance stabilizing transformation, background correction, false discovery rate, normalization, gene expression data
MeSH terms
Algorithms, Humans, Reproducibility of Results, Oligonucleotide Probes, Genome, Oligonucleotide Array Sequence Analysis, Gene Expression Profiling

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review