PRANK Changelog

What's new in PRANK 140110

Feb 19, 2014
  • Search order of external tools changed and bundled programs now preferred. Ancestor inference for translated alignments of DNA sequences. A fix for rare crashes due to overly long branch lengths.

New in PRANK 130820 (Aug 22, 2013)

  • Disabled computation of alignment score for the guide tree alignment. More information about iteration with a user-provided tree. Path to other binaries no more affected by renaming of the program file.

New in PRANK 130708 (Jul 17, 2013)

  • Significant bug fixes :
  • option -F was mistakenly turned off by the new iterative approach
  • without option -F, ancestor reconstruction (and scoring) were incorrect

New in PRANK 121218 (Jan 4, 2013)

  • More detailed information about unmatching names. New option "-prunedata".

New in PRANK 121002 (Oct 6, 2012)

  • Alignment merge now accepts trees such as "(t1:#.#,t2:#.#);". The two-node tree can be provided with "-t=filename" or "-tree=tree-string".

New in PRANK 120716 (Jul 17, 2012)

  • Fixed the translated alignment option that had been broken in recent clean up, and corrected the output order for ancestral sequences.

New in PRANK 120712 (Jul 14, 2012)

  • Bug fixes: the new speed-up with MAFFT often failed with codon sequence data. That has now been fixed. Also, for codon data, the guide tree is now computed from a MAFFT alignment of translated protein sequences.

New in PRANK 120626 (Jul 11, 2012)

  • The new version of PRANK automatically uses programs MAFFT and Exonerate to speed up the alignment. If these programs are available, the analysis starts with a quick MAFFT alignment and estimation of a NJ guide tree from that. Exonerate is used to anchor the pairwise alignments and thus massively reduce the computation time. The anchoring works for DNA, protein and codon alignments, the last one using translation to proteins in the anchoring step. The guide tree estimation and anchoring work automatically and can be disabled with options -nomafft and -noanchors, respectively.
  • In addition to the alignment anchoring, the new version also provides an option to merge two pre-existing alignments. When merging two alignments, PRANK first infers the ancestors and the sequence at the root for each of them and then aligns the two root sequences. The new option is based on a previous (undocumented) function to finish a partially aligned alignment. See the documentation page for details.
  • Also the previous versions of PRANK have been able to infer the ancestral sequences for a new or an existing alignment. However, the output format for this was somewhat unclear and a new option -showanc has now been added with a more straightforward output.
  • Finally, the default options of the program have been slightly altered. The old behaviour and output (e.g., guide tree and xml file) can be achieved with appropriate options. See prank -h for details.