Comparison of different nucleotide or amino acid substitution models
Usage
modelTest(object, tree = NULL, model = NULL, G = TRUE, I = TRUE,
FREQ = FALSE, k = 4, control = pml.control(), RHAS = "gamma", ...,
mt_control = mt.control(crit = "BIC", n_model = 5, n_rhas = 100))
write.modelTest(x, file = "modelTest", digits = 10)Arguments
- object
an object of class phyDat or pml.
- tree
a phylogenetic tree.
- model
a vector containing the substitution models to compare with each other or "all" to test all available models.
- G
logical, TRUE (default) if (discrete) Gamma model should be tested.
- I
logical, TRUE (default) if invariant sites should be tested.
- FREQ
logical, FALSE (default) if TRUE amino acid frequencies will be estimated.
- k
number of rate classes. Can be a list with a vector for each RHAS term.
- control
A list of parameters for controlling the fitting process.
- RHAS
a character vector specifying the rate heterogeneity among sites models. Option are "gamma" for discrete gamma model with equal weights, "variable_gamma" for discrete gamma model with estimated weights, "free_rate" and "gamma_quadrature".
- ...
Further arguments passed to or from other methods.
- mt_control
a list with some options.
- x
an object of class modelTest.
- file
a file name. File endings are added.
- digits
default is 10, i.e. edge length for the bootstrap trees are exported. For digits larger smaller than zero no edge length are exported.
Value
A data.frame containing the log-likelihood, number of estimated parameters, AIC, AICc and BIC all tested models. The data.frame has an attributes "env" which is an environment which contains all the trees, the data and the calls to allow get the estimated models, e.g. as a starting point for further analysis (see example).
Details
modelTest estimates all the specified models for a given tree and
data similar to (Posada and Crandall 1998; Posada 2008)
.
model="3Di" tests the three models for structural alignments mention in Garg and Hochberg (2025) .
References
Garg SG, Hochberg GKA (2025). “A General Substitution Matrix for Structural Phylogenetics.” Molecular Biology and Evolution, 42(6), msaf124. ISSN 1537-1719. doi:10.1093/molbev/msaf124 .
Posada D (2008). “jModelTest: Phylogenetic Model Averaging.” Molecular Biology and Evolution, 25(7), 1253–1256. doi:10.1093/molbev/msn083 .
Posada D, Crandall K (1998). “MODELTEST: testing the model of DNA substitution.” Bioinformatics, 14(9), 817–818.
Burnham, K. P. and Anderson, D. R (2002) Model selection and multimodel inference: a practical information-theoretic approach. 2nd ed. Springer, New York
Darriba D., Taboada G.L., Doallo R and Posada D. (2011) ProtTest 3: fast selection of best-fit models of protein evolution. . Bioinformatics 27: 1164-1165
Puente-Lelievre, Caroline, et al. (2023) "Tertiary-interaction characters enable fast, model-based structural phylogenetics beyond the twilight zone." bioRxiv: 2023-12
Author
Klaus Schliep klaus.schliep@gmail.com
Examples
if (FALSE) { # \dontrun{
data(Laurasiatherian)
mT <- modelTest(Laurasiatherian, model = c("JC", "K80", "HKY", "GTR"))
# Some exploratory data analysis
plot(mT$TL, mT$logLik, xlim=c(3,6.5),
xlab="total tree length", ylab="log likelihood")
text(mT$TL, mT$logLik, labels=mT$Model, pos=4)
fit_GTR_G <- as.pml(mt, "GTR+G(4)")
fit_GTR_VG <- as.pml(mt, "GTR+VG(4)")
fit_GTR_R <- as.pml(mt, "GTR+R(4)")
plotRates(fit_GTR_G)
plotRates(fit_GTR_VWG, append=TRUE, col="red")
plotRates(fit_GTR_R, append=TRUE, col="green")
# extract best model
(best_model <- as.pml(mT))
data(chloroplast)
(mTAA <- modelTest(chloroplast, model=c("JTT", "WAG", "LG")))
# test all available amino acid models
plan(multisession, workers = 2)
(mTAA_all <- modelTest(chloroplast, model="all"))
plan(sequential)
} # }