Table 1.
Models | ω (dN/dS) | ln La | npb | Models Compared | 2Δ(ln L)c | P Valuesd |
Dataset I: 39 sequences (all except vampire bats) | ||||||
A. All branches have one ω | ω = 0.192 | −6223.77 | 78 | |||
B. All branches have the same ω = 1 | ω = 1 | −6485.21 | 77 | B vs. A | 522.88 | 1.0 × 10−115 |
C. Branches in yellow have ω1, branches in blue have ω2, and branches a, b, c, and d each has its own ω | ω1 = 0.183, ω2 = 0.205 | −6211.42 | 84 | |||
D. Branches in yellow have ω1, branches in blue have ω2 = ω1, and branches a, b, c, and d each has its own ω | ω1 = ω2 = 0.196 | −6211.42 | 83 | D vs. C | 0.64 | 0.422 |
E. Each branch has its own ω | Variable ω by branch | −6189.76 | 153 | A vs. E | 68.02 | 0.703 |
Dataset II: 40 sequences (dataset I plus the ancestral sequence of vampire bats) | ||||||
F. All branches have the same ω | ω = 0.211 | −6575.83 | 80 | |||
G. Branch e (ancestral to vampire bats) has ω2 and other branches have ω1 | ω1 = 0.192, ω2 = 0.994 | −6559.21 | 81 | F vs. G | 33.23 | 8.2 × 10−9 |
H. Branch e has ω2 = 1 and other branches have ω1 | ω1 = 0.192, ω2 = 1 | −6223.77 | 79 | H vs. G | 0 | 1 |
Dataset III: 42 sequences (all bats, after removing gaps in pseudogenes) | ||||||
I. Branch e and branches connecting the three vampire bats have ω2 whereas other branches have ω1 | ω1 = 0.193, ω2 = 0.424 | −6158.29 | 85 | |||
J. Branch e has ω3, branches connecting the three vampire bats have ω2, and other branches have ω1 | ω1 = 0.193, ω2 = 0.393, ω3 = 911.06 | −6156.79 | 86 | I vs. J | 3.00 | 0.083 |
The natural logarithm of the likelihood value.
Number of parameters.
Twice the difference in ln L between the two models compared.
P values lower than 0.05 are shown in bold.