Presenza di outlier nei saggi biologici con PCR real-time: L1-norm (LAD) or L2-norm (OLS)?

Real-time PCR is a laboratory technique which is used to amplify and simultaneously quantify the concentration of biological markers based on sequences of nucleic acid (DNA or RNA). We investigated data from an external quality control of real-time PCR and we observed the presence of outliers. The aim of this work is to compare the performance of L2-norm (OLS) regression method with L1-norm (LAD) robust regression method in estimating the nucleic acid concentration in the presence of outliers. For this purpose a simulation study was planned and two different procedures (OLS and LAD) were applied to the simulated data. Both non contaminated and contaminated (with outliers) situations were considered. In the absence of outliers OLS procedure enables obtaining a precise and accurate estimate of the nucleic acid concentration, whereas the presence of outliers leads to a biased estimate, which can be improved by using LAD robust method.