Improving qPCR telomere length assays: Controlling for well position effects increases statistical power
2014-12-21 - resubmission Comparison of telomere length measurements from monochrome multiplex qPCR.docx (78.32Kb)
Figure 4- Correlation coefficients between TL and validation measures in relation to varying the SE exclusion cut points.eps (137.4Kb)
Eisenberg, Dan T. A.
Kuzawa, Christopher W.
Hayes, M. Geoffrey
MetadataShow full item record
Objectives: Telomere length (TL) is commonly measured using quantitative PCR (qPCR). Although easier than the southern blot of terminal restriction fragments (TRF) TL measurement method, one drawback of qPCR is that it introduces greater measurement error and thus reduces the statistical power of analyses. To address a potential source of measurement error, we consider the effect of well position on qPCR TL measurements. Methods: qPCR TL data from 3,638 people run on a Bio-Rad iCycler iQ are reanalyzed here. To evaluate measurement validity, correspondence with TRF, age and between mother and offspring are examined. Results: First, we present evidence for systematic variation in qPCR TL measurements in relation to thermocycler well position. Controlling for these well-position effects consistently improves measurement validity and yields estimated improvements in statistical power equivalent to increasing sample sizes by 16%. We additionally evaluated the linearity of the relationships between telomere and single copy gene control amplicons and between qPCR and TRF measures. We find that, unlike some previous reports, our data exhibit linear relationships. We introduce the standard error in percent, a superior method for quantifying measurement error compared to the commonly used coefficient of variation. Using this measure, we find that excluding samples with high measurement error does not improve measurement validity. Conclusions: Future studies using block-based thermocyclers should consider well position effects. Since additional information can be gleaned from well position corrections, re-running analyses of previous results with well position correction could serve as an independent test of the validity of these results.