Figure 3.

Summary of toxicity category prediction for 24 chemicals using HPS4138 cells

Red dots indicate predicted SVM values of iPS cell data, and filled and open markers indicate true and false predictions, respectively. Black dots and bars indicate the means ± SEM of SVM values for random data. In the tables, the label columns contain prior knowledge regarding whether the chemical shows toxicity (P: positive) or not (N: negative). The SVM column contains the SVM values of the iPS cell data. The expected accuracy (Exp.Acc.) columns contain the prediction accuracy using random data. The graphs below the tables show the probability distribution of the prediction accuracy using random data. Black lines indicate the probability density estimated using the t distribution (degrees of freedom = 9). Black shaded areas represent the upper 5%. Red dashed lines indicate the prediction accuracy using iPS cell data.