Skip to content

Latest commit

 

History

History
109 lines (84 loc) · 5.73 KB

CHANGELOG.md

File metadata and controls

109 lines (84 loc) · 5.73 KB
Raw

  • Version 1.9.2:

    • omit depreciated rU
    • better exception handling

  • Version 1.9.1:

    • Better compatibility with new dendropy versions

  • Version 1.9.0:

    • Important bug fix: Version 1.8.6 had created a bug where for datasets with less than 1000 sequences (N), the default mask_gappy_sites was set to N^2/1000 instead of N/1000.

  • Version 1.8.7:

    • New feature: run_seqtools.py can be run with portions now

  • Version 1.8.6:

    • New feature: --mask-gappy-sites can now be a number below 1.0 to indicate a percentage

  • Version 1.8.5:

    • Bug fix: handling of U in protein when used with Ginsi is fixed

  • Version 1.8.4:

    • Bug fix: handling of U in protein had a bug where U was repeated multiple times if preceded by -.

  • Version 1.8.3:

    • Bug fix: an import was broken. Fixed
    • Add a new --alignment-suffix option

  • Version 1.8.2:

    • Bug fix: python 3 was broken in 1.8.0. Fixed

  • Version 1.8.1:

    • Bug fix: Version 1.8.0 had a spurious commit

  • Version 1.8.0: This is a major change (by Uyen Mai).

    • Default decomposition is changed to minimize the number of subsets given constraints
    • A new way of building the spanning tree is added as an option (MST)
    • Several other new options for decomposition are added

  • Version 1.7.8:

    • Bug Fix: run_pasta_gui.py should work with python2 now

  • Version 1.7.7:

    • Bug Fix: a second attempt at fixing troubles with new dendropy rooting

  • Version 1.7.6:

    • Bug Fix: PASTA used to sometimes fail due to troubles with new dendropy rooting

  • Version 1.7.4:

    • Bug Fix: fixed a bug that eliminated one sequence

  • Versions 1.7.1 -- 1.7.3:

    • Bug Fix: several bug fixes related to the import to python 3

  • Version 1.7.0:

    • Change: Moved to Dendropy 4 and now support python 3

  • Version 1.6.4:

    • Change: Fix the copyright year to 2013-2016

  • Version 1.6.3:

    • Change: Initial alignment now uses multiple cpus when available

  • Version 1.6.2:

    • Bug fix: When errors happen, PASTA stops right away (also checks for empty alignments)
    • Bug fix: RAxML partition file was not updated according to masked sites
    • Bug fix: -gtr was hard-coded for FastTree when model was empty; now -gtr is NOT automatically added when a configuration file is used and this enables using JK model. (is backward incompatible when a configuration file is used with FastTree as the tree estimator and the model is not set for FastTree in the configuration file)

  • Version 1.6.1:

    • Change: run_seqtools.py is redesigned. It does not use actions anymore (backward incompatible).
    • Change: args can be set in configuration files for each tool and these will be passed to respective tools
    • Feature: Muscle is now available as an alignment tool

  • Version 1.6 (is backward incompatible for AA datasets):

    • Change: We use opal now for pairwise mergers regardless of whether the datatype is AA or DNA.

  • Version 1.5.3:

    • Feature: A new script called run_seqtools.py is added that can be used to manipulate alignments
    • Bug fix: For the initial alignment if all the masked sites were at the beginning, an error occurred.

  • Version 1.5.2:

    • Bug fix: Sometimes, max subset size was set to 201. fixed the issue

  • Version 1.5.1:

    • Bug fix: RAxML post alignment step is fixed, and it uses the masked alignment
    • Bug fix: Longest branch decomposition fixed - subsets of size 1 do not fail anymore
    • Change: User set or automatically set subproblem size is forced to be at least 2
    • Bug fix: FastTree is fixed to use (at most) the requested number of cpus
    • Change: FastTree does not seem to be helped with more than 4 cpus. Max it at 4 cpus.
    • Change: Mafft is moved to - Version 7.15
    • Change: Opal is moved to - Version 2.1.2
    • Change: Mafft is passed a --thread option to use multiple threads for starting alignment
    • Change: Mafft-linsi is used for upto 50,000 length

  • Version 1.5 (is backward incompatibile):

    • Major Change --auto option is not necessary anymore. Important settings are picked automatically based on the input file. These automatically picked defaults can be overwritten by the user. If --auto option is provided, it behaves like before (it overwrites the user provided options)
    • Bug fix: Fasttree for DNA was missing -gamma option by default. Added!
    • Change: blind_after_total_iter is set by default to 0, so that all changes are logged as accepted (before it switched to blind on first worse score)
    • Change: pasta local home is now under ~/.pasta instead of ~/.sate
    • Major Change: file path config file is called ~/.pasta/pasta_tool_paths.cfg instead of ~/.sate/sate_tool_paths.cfg
    • Change: If no job name is given, it will be set by default to pastajob instead of satejob
    • Change: PASTA output (and logs) will say PASTA INFO instead of SATe INFO. More generally, the logs will say PASTA instead of SATe.
    • Change: To turn on the debugging, now you need to use PASTA_DEBUG and PASTA_LOGGING_LEVEL=DEBUG instead of SATE_DEBUG and SATE_LOGGING_LEVEL
    • Change: PASTA code changed such that in most places it uses pasta instead of sate. Note: the config file still has a section named [sate]. This was not renamed to maintain compatibility.

  • Version 1.4.1:

    • Bug fix: GUI was broken in 1.4.0
    • Change: Disable multi-locus checkbox. Multi-locus is not supported in PASTA.

  • Version 1.4.0:

    • Major Change: Uses HMMER to build the initial alignment if one is not given
    • Change: Initial tree also uses a masked alignment