Following the overview plot comes an ordered list of the most
probable sites for 3'-end processing. As shown below for *CYC1*,
the list includes the position, e-value, and DSM score. The
e-value is the expected number of sites with this score or
higher given the length of the query sequence. The list is
cut off at sites with DSM >= 3.0, which corresponds roughly
to a probability of 0.008 of occurrence in random sequence.

The probability of occurrence in random sequence was obtained
empirically through the analysis of 2 x 10^{5} nucleotides
of sequence that were generated with 2nd order statistics (preserving
nucleotide, di-nucleotide, and tri-nucleotide frequencies)
from yeast transcripts (including 5'UTR, CDS, and 3'UTR). The
e-value is obtained by multiplying the probability of occurrence
by the length of the query sequence.