Figure 1

The number of false positives versus the number of random matches. The mean number of matches of a pattern in N (N = 1000) outcomes of a biological sequence dataset randomization (λ) versus the number of false positives (FP) on the biological dataset. Each point corresponds to a PROSITE pattern. The two non-straight lines over and below the bisector delimitate the 95% confidence intervals around the line λ = FP and divide the plan into three regions: I, II and III. Fig. 1a and b (which is a zoom of the square area of Fig. 1a) display sprot100 data, whereas Fig. 1c and d represent human100 and yeast100 data, respectively.