rxCOV is a quantitative metric for assessing immunoassay analyte fidelity

Immunoassay based bioanalytical measurements are widely used in a variety of biomedical research and clinical settings. In these settings they are assumed to faithfully represent the experimental conditions being tested and the sample groups being compared. Although significant technical advances have been made in improving sensitivity and quality of the measurements, currently no metrics exist that objectively quantify the fidelity of the measured analytes with respect to noise associated with the specific assay. Here we introduce ratio of cross-coefficient-of-variation (rxCOV), a fidelity metric for objectively assessing immunoassay analyte measurement quality when comparing its differential expression between different sample groups or experimental conditions. We derive the metric from first principles and establish its feasibility and applicability using simulated and experimental data. We show that rxCOV assesses fidelity independent of statistical significance, and importantly, identifies when latter is meaningful. We also discuss its importance in the context of averaging experimental replicates for increasing signal to noise ratio. Finally, we demonstrate its application in a Lynch Syndrome case study. We conclude by discussing its applicability to multiplexed immunoassays, other biosensing assays, and to paired and unpaired data. We anticipate rxCOV to be adopted as a simple and easy-to-use fidelity metric for performing robust and reproducible biomedical research.

Automated summary

Immunoassay based bioanalytical measurements are widely used in biomedical research and clinical applications. However, there is currently no objective metric to assess the accuracy of these measurements. This study introduces a fidelity metric called ratio of cross-coefficient-of-variation (rxCOV) to objectively evaluate the measurement quality of immunoassay analytes in comparing different sample groups or experimental conditions. The feasibility and applicability of rxCOV are demonstrated using simulated and experimental data, showing its independence from statistical significance and ability to identify meaningful differences. Its importance in increasing signal to noise ratio and its potential applications are also discussed. The adoption of rxCOV as a simple and easy-to-use fidelity metric for robust and reproducible biomedical research is anticipated.