MetaQC: Objective Quality Control and Inclusion/Exclusion Criteria for Genomic Meta-Analysis

Introduction

Genomic meta-analyses for combining microarray studies have been widely applied to increase statistical power and to validate results from individual studies. Currently, the inclusion/exclusion criteria mostly depend on ad-hoc expert opinion or naïve decision by sample size or array platform. No objective quality assessment is available.

MetaQC implements our proposed quantitative quality control measures:

• (1) internal homogeneity of co-expression structure among studies (internal quality control; IQC)
• (2) external consistency of co-expression structure correlating with pathway database (external quality control; EQC)
• (3) accuracy of differentially expressed gene detection (accuracy quality control; AQCg) or pathway identification (AQCp)
• (4) consistency of differential expression ranking in genes (consistency quality control; CQCg) or pathways (CQCp).

For each quality control index, the p-values from statistical hypothesis testing are minus log transformed and PCA biplots were applied to assist visualization and decision. Results generate systematic suggestions to exclude problematic studies in microarray meta-analysis and potentially can be extended to GWAS or other types of genomic meta-analysis. The identified problematic studies can be scrutinized to identify technical and biological causes (e.g. sample size, platform, tissue collection, preprocessing etc) of their bad quality or irreproducibility for final inclusion/exclusion decision.

Installation

To install this package, run:

    source("http://bioconductor.org/biocLite.R")
    biocLite('MetaQC')

Examples

library(MetaQC)

## Toy Example
data(brain) #already hugely filtered

#Two default gmt files are automatically downloaded, 
#otherwise it is required to locate it correctly.
#Refer to http://www.broadinstitute.org/gsea/downloads.jsp
brainQC <- MetaQC(brain, "c2.cp.biocarta.v3.0.symbols.gmt", filterGenes=FALSE, verbose=TRUE)
#B is recommended to be >= 1e4 in real application					
runQC(brainQC, B=1e2, fileForCQCp="c2.all.v3.0.symbols.gmt") 
brainQC
plot(brainQC)

## For parallel computation with only 2 cores
brainQC <- MetaQC(brain, "c2.cp.biocarta.v3.0.symbols.gmt", filterGenes=FALSE, verbose=TRUE, isParallel=TRUE, nCores=2)
#B is recommended to be >= 1e4 in real application
runQC(brainQC, B=1e2, fileForCQCp="c2.all.v3.0.symbols.gmt") 
plot(brainQC)

## For parallel computation with half cores
## In windows, only 3 cores are used if not specified explicitly
brainQC <- MetaQC(brain, "c2.cp.biocarta.v3.0.symbols.gmt", filterGenes=FALSE, verbose=TRUE, isParallel=TRUE)
#B is recommended to be >= 1e4 in real application					
runQC(brainQC, B=1e2, fileForCQCp="c2.all.v3.0.symbols.gmt") 
plot(brainQC)

## Real Example which is used in the paper
#download the brainFull file from https://github.com/downloads/donkang75/MetaQC/brainFull.rda
load("brainFull.rda")
brainQC <- MetaQC(brainFull, "c2.cp.biocarta.v3.0.symbols.gmt", filterGenes=TRUE, verbose=TRUE, isParallel=TRUE)
runQC(brainQC, B=1e4, fileForCQCp="c2.all.v3.0.symbols.gmt") #B was 1e5 in the paper 
plot(brainQC)

## Survival Data Example
#download Breast data 
#from https://github.com/downloads/donkang75/MetaQC/Breast.rda
load("Breast.rda")
breastQC <- MetaQC(Breast, "c2.cp.biocarta.v3.0.symbols.gmt", filterGenes=FALSE, verbose=TRUE, isParallel=TRUE, resp.type="Survival")
runQC(breastQC, B=1e4, fileForCQCp="c2.all.v3.0.symbols.gmt") 
breastQC
plot(breastQC)

References

Dongwan D. Kang, Etienne Sibille, Naftali Kaminski, and George C. Tseng. (Nucleic Acids Res. 2012) MetaQC: Objective Quality Control and Inclusion/Exclusion Criteria for Genomic Meta-Analysis. (doi: 10.1093/nar/gkr1071)