Bui Quang Minh
We are happy to announce version 1.4.3, which fixes various problems improving software stability.
New features:
* Better parameter estimates for I+G model.
* More flexible mixture models for model testing with `-madd` option (requested by David Kerk).
* The protein mixture model `CF4` of Wang et al. (2008) now includes Gamma rate heterogeneity by default.
* Support invariable sites plus FreeRate [+I+R] model (requested by Lars Jermiin).
* New option `--sequential` to read sequential phylip alignment file format.
Bug fixes:
* Failture too estimate too extreme GTR rate parameters (reported by Stephen Crotty).
* Bug with likelihood scaling for constant sites under invariable site [+I] model (reported Remi Denise).
* Crash with optimizing codon model parameters (reported by Xiaofan Zhou).
* Redundant codon models for model selection (reported by Xiaofan Zhou).
* Segfault caused by unaligned memory for partition model with binary data (reported by Marek L. Borowiec).
* Crash with `-wbtl` option (reported by Teo).
* Crash with `-mtree` for partition finding (reported by a web user).
* A rare bug in NNI hill-climbing search.
* A bug in printing .rate file via -wsr option (reported by Tim McInerney).
* Several other minor issues.