Ultrafast (partition) model selection for phylogenomic alignments.
Higher accuracy in tree reconstruction and bootstrap with more thorough nearest neighbor interchange enabled via -nni5 option (optimizing 5 branches around NNI). This comes at the trade-off of c.a. 2X longer running time.
Introduction of joint and proportional partition models to reduce the number of parameters in case of model overfitting (EXPERIMENTAL)
Introduction of gene-resampling and gene-and-site resampling for the bootstrap on multi-gene alignments.
Introduction of epsilon for ultrafast bootstrap: trees similar RELL log-likelihoods will be chosen at random to break tie. This helps to reduce over-optimistic supports in case of polytomies.
Tree topology tests (BP,KH,SH,ELW,WKH, and WSH tests via RELL method).
Parallel OpenMP version for multi-core CPUs.
New implementation of model selection that works for DNA, amino-acid, and binary models.