Fig. 1.

Distribution of quality. a Intra-experimental quality heterogeneity. When many data records are produced by a single experiment, the quality of these records can vary. The majority of the records are good, but a tail of poorer quality is generated due to multiple factors. b Inter-experimental quality heterogeneity is shown as the difference of quality distributions. c When the quality of a small fraction of experiments is apparently poorer than the majority, a particular cause is strongly suspected, which gives good reason to remove the fraction of the experiments by setting a threshold to discriminate these experiments. d When the quality distribution is smooth, the selection of a threshold value can be arbitrary