Comparison of the top 20 8-mers calculated using six different methods.

8-mers derived from the region 500 bases upstream of all yeast ORFs were ranked by their relative abundance as calculated by several background models. The top 20 8-mers ranked by decreasing fold enrichment scores from (a) our novel background model, the “Ak-1 method” were compared to those from the previously described C0/C1 method (b), from Markov models of order 4 sorted by Z-scores (c), from Markov models of order 4 sorted by Binomial probabilities (d), from Markov models of order 3 ranked by Z-scores (e) and, Markov models of order 3 ranked by binomial probabilities (f). The list of 20 8-mers derived by applying our method was chosen as the reference against which the other lists were compared. 8-mers in each list that are identical to the ones in the reference list are marked with a superscript number which shows its rank in the reference column (a), while repeat-rich elements are shown italicized. The nine 8-mers identified uniquely by our method are highlighted in bold (a). The average GC content of each list is shown in the bottom row of the table.