Comparison of traditional quantitative pcr with computer quantitation

Background

Difference between traditional quantitative pcr with computer-based quantitation algorithm for cmv. Traditional viral quantitation using cycle threshold (Ct) analysis is dependent on generating an accurate, valid calibration curve that is stable over time. The relationship between the Ct values obtained and the expected concentration for each calibrator is used to establish a formula to quantify viral concentrations in patient samples. Additionally, a fluorescence threshold is set to compare amplification cycles as an indirect measure of the starting DNA concentration. DNA Software, Inc. (Ann Arbor, MI) has developed a curve analysis algorithm (qPCR CopyCount) based on Poisson distribution that analyzes qPCR data to obtain an absolute DNA copy number without the need for a standard curve or use of Ct values. Our objectives were to compare the performance of qPCR CopyCount to traditional qPCR (Ct) for the quantitation of CMV from plasma and evaluate its feasibility for daily clinical use in a CMV plasma quantitation assay.

Methods

Fluorescence data from archived runs of our CMV assay, using Abbott ASR reagents on the Abbott m2000 system, were analyzed using CopyCount algorithm. CopyCount quantification was compared to original data using a calibration curve and Ct values. Nucleic acid quantitative standards (Qiagen), calibrators traceable to international units (CMVtc panel, Acrometrix), and patients were compared for linearity, precision, and accuracy.

Results

CopyCount results had a linear relationship with nucleic acid quantitative standard assigned values (r2>0.99; bias of log10 0.2 copies per reaction) and patient results by Ct method (r2>0.99). Raw CopyCount values do not account for extraction efficiency or concentration of the eluate, introducing a systematic bias between methods (bias -1.2 copies/ml). Evaluation of the CopyCount values from the extracted CMVtc panel demonstrated a linear relationship (r2>0.98) and provided a correction factor for extraction efficiency. Using this correction, the relationship between CopyCount and patient values remained linear (r2>0.99) with a reduced bias of log10 0.08 IU/ml. Precision of quantitative standard results (4 standards in duplicate over 10 runs) was a total standard deviation range of 0.05-0.15 log10 copies/rxn for CopyCount vs 0.02-0.16 log10 copies/ml for the current Ct method.

Conclusions

The CopyCount method performed similarly to the conventional Ct method for quantitation of CMV from patient samples and quantitative standards with excellent linearity and comparable precision. A correction factor or formula is required to DNA concentration in the PCR reaction to the concentration in the original specimen. A one-time calibration, using 1.5 copies per reaction as recommended by the software company to optimize the curve analysis formula, may further improve the accuracy of quantitation.