Evaluation of normalization methods for two-channel microRNA microarrays

Yingdong Zhao, Ena Wang, Hui Liu, Melissa Rotunno, Jill Koshiol, Francesco M. Marincola, Maria T. Landi, Lisa M. McShane

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

Background: MiR 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.Methods: We 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.Results: Lowess 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.Conclusions: Researchers 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.

Original languageEnglish
Article number69
JournalJournal of Translational Medicine
Volume8
DOIs
Publication statusPublished - 21 Jul 2010
Externally publishedYes

Fingerprint

Microarrays
MicroRNAs
Gene expression
Cells
Gene Expression
Research Personnel
Cell Line

ASJC Scopus subject areas

  • Medicine(all)
  • Biochemistry, Genetics and Molecular Biology(all)

Cite this

Zhao, Y., Wang, E., Liu, H., Rotunno, M., Koshiol, J., Marincola, F. M., ... McShane, L. M. (2010). Evaluation of normalization methods for two-channel microRNA microarrays. Journal of Translational Medicine, 8, [69]. https://doi.org/10.1186/1479-5876-8-69

Evaluation of normalization methods for two-channel microRNA microarrays. / Zhao, Yingdong; Wang, Ena; Liu, Hui; Rotunno, Melissa; Koshiol, Jill; Marincola, Francesco M.; Landi, Maria T.; McShane, Lisa M.

In: Journal of Translational Medicine, Vol. 8, 69, 21.07.2010.

Research output: Contribution to journalArticle

Zhao, Y, Wang, E, Liu, H, Rotunno, M, Koshiol, J, Marincola, FM, Landi, MT & McShane, LM 2010, 'Evaluation of normalization methods for two-channel microRNA microarrays', Journal of Translational Medicine, vol. 8, 69. https://doi.org/10.1186/1479-5876-8-69
Zhao, Yingdong ; Wang, Ena ; Liu, Hui ; Rotunno, Melissa ; Koshiol, Jill ; Marincola, Francesco M. ; Landi, Maria T. ; McShane, Lisa M. / Evaluation of normalization methods for two-channel microRNA microarrays. In: Journal of Translational Medicine. 2010 ; Vol. 8.
@article{98fe985e63644646abfbe0aa8c159b13,
title = "Evaluation of normalization methods for two-channel microRNA microarrays",
abstract = "Background: MiR 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.Methods: We 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.Results: Lowess 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.Conclusions: Researchers 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.",
author = "Yingdong Zhao and Ena Wang and Hui Liu and Melissa Rotunno and Jill Koshiol and Marincola, {Francesco M.} and Landi, {Maria T.} and McShane, {Lisa M.}",
year = "2010",
month = "7",
day = "21",
doi = "10.1186/1479-5876-8-69",
language = "English",
volume = "8",
journal = "Journal of Translational Medicine",
issn = "1479-5876",
publisher = "BioMed Central",

}

TY - JOUR

T1 - Evaluation of normalization methods for two-channel microRNA microarrays

AU - Zhao, Yingdong

AU - Wang, Ena

AU - Liu, Hui

AU - Rotunno, Melissa

AU - Koshiol, Jill

AU - Marincola, Francesco M.

AU - Landi, Maria T.

AU - McShane, Lisa M.

PY - 2010/7/21

Y1 - 2010/7/21

N2 - Background: MiR 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.Methods: We 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.Results: Lowess 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.Conclusions: Researchers 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.

AB - Background: MiR 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.Methods: We 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.Results: Lowess 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.Conclusions: Researchers 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.

UR - http://www.scopus.com/inward/record.url?scp=77954670531&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77954670531&partnerID=8YFLogxK

U2 - 10.1186/1479-5876-8-69

DO - 10.1186/1479-5876-8-69

M3 - Article

C2 - 20663154

AN - SCOPUS:77954670531

VL - 8

JO - Journal of Translational Medicine

JF - Journal of Translational Medicine

SN - 1479-5876

M1 - 69

ER -