5 months ago by
Wageningen University, Wageningen, the Netherlands
In practice we almost never look at the values of these metrics as such, but rather look for samples that deviate from the majority. We then check whether e.g. pseudo-images indicate the presence of stains, or that some samples e.g. belong to a different tissue (e.g. liver versus intestine). In case of the later deviations are very likely because of a 'biological' reason.
Regarding the NUSE box plots; for us this is the most sensitive QC parameter, and we use as rule of thumb that a sample passes that parameter if the corresponding 'box' (i.e. first and third quartiles) still 'touches' the 1.00 line. If that's not the case, we check the other QC parameters and sample source if there is a technical problem or not.
I found these papers very informative regarding interpretation of the various QC metrics:
Brettschneider et al. Quality Assessment for Short Oligonucleotide Microarray Data. Technometrics, 2008;50:3, 241-264. DOI.
Heber & Sick. Quality assessment of Affymetrix GeneChip data. OMICS. 2006;10(3):358-68. DOI.
McCall et al. Assessing affymetrix GeneChip microarray quality. BMC Bioinformatics. 2011;12:137. DOI.
Rosikiewicz & Robinson-Rechavi. IQRray, a new method for Affymetrix microarray quality control, and the homologous organ conservation score, a new benchmark method for quality control metrics. Bioinformatics. 2014;30(10):1392-9. DOI.