Skip to main content

Table 2 Comparison of running time and used space of SUFPREF and AHOPRO programs for PSSM-based patterns of length 12

From: Analysis of pattern overlaps and exact computation of P-values of pattern occurrences numbers: case of Hidden Markov Models

Experiments parameters Time Space
Pattern Fraction() Prob Distrib SufPref AhoPro Aho/SP SufPref AhoPro Aho/SP
PSSM(12,9.63) 0.00001 Bernoulli 0.02 0.37 20.39 0.44 0.59 1.36
PSSM(12,8.69) 0.00003 Bernoulli 0.03 0.90 32.00 0.5 0.97 1.94
PSSM(12,7.41) 0.0001 Bernoulli 0.07 2.60 37.64 0.69 1.88 2.74
PSSM(12,5.89) 0.0003 Bernoulli 0.27 7.64 28.10 1.21 4.97 4.11
PSSM(12,4.01) 0.001 Bernoulli 1.27 26.15 20.61 3.01 15.28 5.07
PSSM(12,2.04) 0.003 Bernoulli 4.99 78.37 15.70 7.75 42.61 5.50
PSSM(12,9.63) 0.00001 Markov 0.03 0.38 15.12 0.47 0.62 1.32
PSSM(12,8.69) 0.00003 Markov 0.05 0.91 18.65 0.53 0.97 1.84
PSSM(12,7.41) 0.0001 Markov 0.11 2.64 23.13 0.71 1.91 2.67
PSSM(12,5.89) 0.0003 Markov 0.41 7.74 18.78 1.24 5.02 4.04
PSSM(12,4.01) 0.001 Markov 1.77 26.50 14.95 3.04 15.31 5.04
PSSM(12,2.04) 0.003 Markov 6.67 79.25 11.88 8.36 42.65 4.94
  1. See Table 1 for the general information on the patterns. The intermediate values of Fraction ( ) (0.003, 0.0003, etc. instead of more common 0.005, 0.0005, etc.) were chosen to obtain more homogeneous log-scale.