Samples that fail quality control in several panels most certainly do so due to sample properties. During quality control all data is also evaluated for position effects on the plates, e.g. if columns or rows of samples deviate from the remaining samples. As we use multichannel pipettes for most steps of the analysis, this will likely pick up most operator errors. If such a difference is found, it is assumed that samples are flagged due to operator error and all flagged samples should be rerun.