Table 1.
Model with positive selectiona | Null modela | LRTb | P-value | |||||
---|---|---|---|---|---|---|---|---|
Model | Log-likelihood | Parametersc | Positively selected sitesd | Model | Log-likelihood | Parametersc | 2L | |
Analysis for positively selected sites common for C3 and C4 clades | ||||||||
M2a | –6741.7 | κ=2.94, p0=0.82, ω0=0.07, ps=0.003, ωS=3.24 | None | M1a | –6742.34 | κ=2.92, p0=0.82, ω0=0.07 | 1.28 | 0.5273 |
M8 | –6740.28 | κ=2.89, p0=0.98, p=0.24, q=1.0, ωS=1.93 | None | M8a | –6743.02 | κ=2.86, p0=0.96, p=0.19, q=0.7 | 5.48 | 0.0192 |
Analysis for positive selection along branches leading to C4 clades | ||||||||
A | –6729.78 | κ=2.95, p0=0.81, ω0=0.08, ps=0.04, ωS=4.2 | 733, 868 | A1 | –6737.24 | κ=2.89, p0=0.77, ω0=0.06 | 14.9 | 0.0001 |
Analysis for positive selection along branches leading to Kranz C4 clades | ||||||||
A | –6732.25 | κ=2.94, p0=0.77,ω0=0.06, ps=0.06, ωS=3.64 | 485, 519, 735 | A1 | –6735.67 | κ=2.91, p0=0.71, ω0=0.06 | 6.8 | 0.0091 |
Analysis for positive selection along branches leading to single-cell C4 clades | ||||||||
A | –6742.34 | κ=2.92, p0=0.82, ω0=0.07, ps=0.00, ωS=NA | None | A1 | –6742.34 | κ=2.92, p0=0.82, ω0=0.07 | 0 | 1 |
Analysis for positive selection along all C4 branches | ||||||||
A | –6702.46 | κ=2.91, p0=0.74, ω0=0.05, ps=0.18, ωS=1.5 | 480, 513, 627, 662, 695, 707, 733, 744, 794, 863, 868, 880, 931 | A1 | –6705.34 | κ=2.86, p0=0.69, ω0=0.04 | 5.08 | 0.0242 |
a M1a (nearly neutral), M2a (positive selection), M8a (beta and ω=1), and M8 (beta and ω) are PAML site models; A1 and A are PAML branch-site models.
b LRT is the likelihood ratio test; 2L is twice the difference of model log-likelihoods.
c κ is the transition/transversion rate ratio; ω is the dN/dS ratio; ωs is the dN/dS ratio in a class under putative positive selection; p0 and ps are the proportion of codons with ω<1 and ω>1, respectively; p and q are parameters of beta distribution in the range (0, 1).
d Sites listed are those at which positive selection is detected at the significance level of >95%, or >99% in bold italics.