Figure 1A–F. Distribution of PWM model prediction scores of positive example sequences and outlier detection thresholds calculated by three different statistical methods for nine representative OGs. Each individual plot shows the histogram of prediction scores for positive examples associated with a particular orthologous group (OG), using the PWM prediction model (see also Table 1, data taken from ref. 1). The dashed red line represents the scaled density function of the normal distribution whose parameters have been estimated from the prediction scores. The solid red vertical corresponds to the mean of the distribution. The purple, green and black vertical lines represent the rejection thresholds for putative false positive examples corresponding to the three different statistical methods (see text). In the upper left corner of each plot the values of the thresholds are provided along with the respective number of rejected examples in parentheses. The following OGs are shown: (A) all sequences; (B) Acyl-CoA oxidase isoform 1 (ACX1); (C) Alanine (serine)-glyoxylate amiontransferase (AGT); (D) Acetyltransferase (ATF); (E) Quinone oxidoreductase (BSMDR); (F) Glutathione S-transferase isoform theta 1 (GSTT1); (G) Hydroxypyruvate reductase (HPR); (H) Malate synthase (MLS); (I) Short-chain dehydrogenase-reductase B/2,4-Dienoyl-CoA reductase (SDR-b/DECR); (J) sterol carrier protein isoform 2 (SCP2).