Skip to main content
. 2020 Mar 28;9(4):64. doi: 10.3390/biology9040064

Table 1.

Log likelihood and AICc values a for empirical models and the GTR model optimized for exposed and buried sites. The GTR model had the best fit (note bold AICc value).

Structural Subset Model T2 b T3 b Cni+Bil c Cni+Pla c lnL AICc
Exposed GTR 92 - - 98 −1,212,232.985 2,424,956.474
LG 100 - - 100 −1,222,489.017 2,445,088.162
WAG 87 - - 100 −1,225,898.553 2,451,907.235
VT 96 - - 98 −1,225,672.633 2,451,455.395
rtREV 90 - - 98 −1,222,733.929 2,445,577.986
JTTDCMUT 94 - - 98 −1,229,028.451 2,458,167.031
Buried GTR - 82 64 - −1,045,694.924 2,091,880.072
LG - 87 62 - −1,050,022.577 2,100,155.267
WAG - 85 58 - −1,054,829.256 2,109,768.626
VT - 81 61 - −1,059,148.671 2,118,407.455
rtREV - 89 52 - −1,054,432.123 2,108,974.359
JTTDCMUT - 81 - 54 −1,061,103.295 2,122,316.704

a AICc was calculated using: AICc = 2k − 2(lnL) + (2k2 + 2k) / (nk − 1) where k is the number of parameters and n is the number of sites. This is the formula for the AICc that is typically used in phylogenetics. b RaxML bootstrap support for T2 or T3. c RAxML bootstrap support for the cnidaria+bilateria clade or the cnidaria + placozoa clade.