Assessment of the algorithm and PASS program. A. The relationship of sensitivity (Sn), specificity (Sp), and threshold. Threshold is a selectable standard in determining whether a poly(A) site is next to a nucleotide or not. It is also measured as a score for each nucleotide of an individual sequence. The higher the threshold, the better the probability that a nucleotide is a poly(A) site. Sn _0, Sn _3, and Sn _10 represent the distance between the prediction site and the validated site to be 0, 3, and 10 nucleotides, respectively. Random 8K, a randomly generated 8000 sequence dataset based on the 2nd order distribution of trinucleotide in the 8K dataset. Coding Seqs, 8000 coding sequences from Arabidopsis (downloaded from TAIR). Intron (8000 sequences) and 5'-UTR (974 sequences) datasets are also from Arabidopsis. B. The average prediction scores of the 8K dataset and other control datasets as in A. The authenticated poly(A) site at location 301 is as mark by a red triangle. C. Distribution of scores in the 8K dataset. The distribution of all other sites (except position 301) is presented as average scores of all these sites. The scores at the 301 position of each of the sequences were counted and their distribution of them is presented.