Skip to main content
. 2016 Sep 27;17:397. doi: 10.1186/s12859-016-1167-6

Table 1.

Perturbation analysis on local PWA probabilities in case (iv)

(∆L A, ∆L D) 0.01 indels/site 0.04 indels/site 0.1 indels/site 0.2 indels/site
(1, 1) 0.003 0.010 0.024 0.045
(3, 1) 0.021 0.084 0.204 0.393
(3, 3) 0.042 0.166 0.402 0.768
(5, 5) 0.073 0.283 0.672 1.256
(10, 1) 0.064 0.246 0.572 1.013
(10, 10) 0.149 0.561 1.292 2.288
(25, 1) 0.151 0.547 1.112 1.541
(25, 4) 0.198 0.723 1.519 2.234
(30, 10) 0.288 1.038 2.164 3.072
(100, 1) 0.537 1.333 1.507 1.574
(100, 3) 0.607 1.593 1.894 2.033
(300, 1) 1.165 1.394 1.427 1.527

Each cell shows the ratio of the total next-parsimonious contribution to the total parsimonious contribution, when there are ∆L A ancestral sites and ∆L D descendant sites in between the PASs. Each column is labeled with the expected number of indels per site ((λ I + λ D)(t F − t I)). See section M1 of Methods for the parameter setting. Because of the symmetry between probabilities under the time reversal, the ratio for (ΔL A, ΔL D) = (L 1, L 2) is identical to that for (ΔL A, ΔL D) = (L 2, L 1) when λ I = λ D. Thus we only showed the results for ΔL A ≥ ΔL D. The ratios that are less than 0.5 are shown in boldface. This table is identical to Table 2 of [43]