Assessment of precision and concordance of quantitative mitochondrial DNA assays: a collaborative international quality assurance study

Background: A number of international research groups have developed DNA quantitation assays in order to investigate the role of mitochondrial DNA depletion in anti-retroviral therapy-induced toxicities. Objectives: A collaborative study was undertaken to evaluate intra-assay precision and between laboratory concordance of measurements of mitochondrial DNA quantity, as a component of a comprehensive quality assurance project. Study design: Four laboratories were asked to measure and report mitochondrial DNA and nuclear DNA genome copy number, as well as mitochondrial DNA copy number/cell, for 17 coded aliquots of DNA derived from serial dilutions of pooled DNA from a lymphoblastoid cell line. Samples included masked replicates and five standards. All samples had similar mitochondrial DNA/nuclear DNA ratios. Precision within laboratories was assessed by determining the coefficient of variation of replicates. Concordance between laboratories was assessed by determining the average coefficient of variation of the mean replicate values for each sample. The effect of standardising the assay for these three measurements was also assessed for laboratories A, B and C. Results: Measurements of mitochondrial DNA and nuclear DNA content for replicate samples varied by an average of less than 6% (based on log 10 values, 72% non-logged values), and measurements of mitochondrial DNA/cell for replicates varied by less than 12% (based on log 10 values, 32% non-logged values), with no improvement of precision after standardisation. Standardisation did significantly improve the concordance of results for measurements of mitochondrial DNA content and mitochondrial DNA/cell. Non-standardised measurements of mitochondrial DNA content for the same sample set varied by 19% between laboratories (based on log 10 values, 96% non-logged values), and after standardisation results varied by less than 3% (based on log 10 values, 54% non-logged values). There was no significant improvement for concordance of measures of nuclear DNA content after standardisation, with results varying by 4.56% between laboratories (based on log 10 values, 45% non-logged values) before standardisation, and by 2.49% (based on log 10 values, 50% non-logged values) after standardisation. Derived values of mitochondrial DNA/cell varied between laboratories by an average of 91% (non-logged, 56% log 10 values) before and by 56% (non-logged, 13% log 10 values) after standardisation. Conclusion: All assays demonst