Skip to main content
. 2009 May 8;37(Web Server issue):W363–W368. doi: 10.1093/nar/gkp299

Figure 1.

Figure 1.

Score distribution for the non-redundant set of all ∼30 000 SwissProt proteins shorter than 150 amino acids. The proteins that were used for the classifier positive training set were excluded. Positive and secured prediction scores are (P2–P3) assigned to the tail of the distribution with a mean score >0.2. The intermediate levels of score ranging between –0.2 and 0.2 are considered probable/possible toxin-like. More refined positive prediction confidence (marked as P3, P2 and P1) is defined according to the scale of the SD relative to the mean score. The negative prediction (N) is associated with a mean score <–0.2 that includes the large Gaussian-like distribution for the vast majority of the proteins.