Release 2.2.7 ============= - [BUG FIX] More intuitive behavior for the maximum subproblem option from the GUI. There are two controls for the maximum subproblem used during tree decomposition, frac and size. Both of these have default values, and the larger of the two is used to determine the maximum subset size during decomposition. This caused confusion when using the GUI, because the selected option was ignored if it was smaller than the default for the alternative option. The GUI now explicitly configures BOTH options as per the user's radio-button selection, so that what the user sees is what they get. In short, always check the configuration file reported by SATe! This will tell you exactly what SATe is doing. Release 2.2.6 ============= - [CHANGES] RNA sequence data is now fully supported. There was some inconsistency in previous versions as to whether RNA data was allowed to run (under some versions and configurations it was allowed to run as DNA data). This version internally converts RNA input to DNA and then converts the final results back to RNA. - [CHANGES] Updated the alignment length check that determines which MAFFT algorithm to use for alignment. Previously this checked the alignment length, it now uses the maximum unaligned sequence length. - [CHANGES] Better error messaging when input sequence labels conflict with tip labels of a user-provided starting tree. The error message now reports the problem and lists all of the mismatched labels. - [CHANGES] This version gives explicit warning messages when the user specifies a max subproblem argument that is equal to or greater than the number of tips. Under these conditions, the SATe algorithm is not invoked. The warning makes this more clear to the user. - [CHANGES] Many tests added to the test suite. Release 2.2.5 ============= - [BUG FIX] Fixed directory cleanup error messages introduced with release 2.2.4. SATe can now remove temporary directories without these annoying error messages. This bug did not affect analyses. - [BUG FIX] Fixed tree-decomposition recursion bug introduced with release 2.2.4. During some recursions, multiple sate aligner jobs would attempt to create the same temp directory, causing the analysis to exit with an error. This release corrects this problem. Any results obtained from 2.2.4 should not be effected by the bug, because this bug always resulted in a crash. Release 2.2.4 ============= - [BUG FIX] Fixed temp directory path depth problem in Windows when using "longest" decomposition. Previous versions created a nested temp file directory that mirrored the tree decomposition. In Windows, if the tree had enough long terminal branches, the temp files got too long for the OS. The temp file structure has been changed to avoid this problem. - [BUG FIX] Fixed problem with too many open files when multi-locus analyses had many loci. Release 2.2.3 ============= - [CHANGES] MUSCLE is now the default merger tool. - [CHANGES] Centroid is now the default decomposition strategy. - [CHANGES] More sample data and output files bundled. Release 2.2.2 ============= - [BUG FIX] Previous version had used the "Centroid" decomposition even when the user specified "Longest." - [BUG FIX] Multiple issues dealing the handling very small subsets of taxa. These arise frequently in the "longest" edge decomposition, but are rare when using centroid decomposition. - [BUG FIX] Input sequences with all gaps no longer create problems (this can reasonably occur in multi-locus runs). - [CHANGES] FastTree binary in Mac bundle is compiled for more architectures. - [CHANGES] Updated documentation. - [CHANGES] RAxML user-specified arguments implemented in CLI. - [CHANGES] Post-processing RAxML search now uses start_tree_search_from_current boolean to determine whether to use the tree resulting from the SATe algorithm as starting tree. - [CHANGES] Updated bundled data. Release 2.2.1 ============= - [BUG FIX] Minor fix to initial GUI display. Initial "Model" options now correspond to FastTree options. - [CHANGES] updated documentation in README and FAQ. Release 2.2.0 ============= - [BUG FIX] No longer uses 'N' for missing taxa in multi-locus mode, which caused problems with multi-locus amino acid analyses with missing taxa. - [CHANGES] Fasttree is now bundled with SATe. - [CHANGES] New option to automatically set defaults based on size of dataset. - [CHANGES] Several default settings changed. Release 2.1.0 ============= - [BUG FIX] No longer write out duplicate sequence entries in final alignment output. - [CHANGES] Do not clean up files and directories in temporary working directory not created by SATe. - [BUG FIX] Check for gamma score in output file before using value in log file. Release 2.0.3 ============= - Fix for not terminating correctly in SATE-II-fast mode. - Allow specification of memory to be allocated to JVM when running Opal [NOTE: default memory allocation is now 1024MB instead of 2048MB]. - Default to SATE-II-fast mode instead of SATE-II-ML. - SATE-II-ML defaults to 10 iterations instead of 100. Release 2.0.2b ============== - Fix for multilocus analysis with non-equal leaf-sets. - Fix for allowing SATe application bundle to be run from directories with non-alphanumeric characters in path. Release 2.0.1b ============== - Support for multilocus analysis. - `pairalign` (MAFFT) is not run if sequences are > 10K characters. This configuration has not been benchmarked.