New features in the latest version of PAGAN
Fixed a serious bug in query placement of independent sequences: previously
substrings (insertions) may have been left out of the resulting alignment.
Also fixed the branch length computation for the new added sequences.
PAGAN code is now hosted at github: changed the update check to reflect that.
Using now explicit data types for Exonerate. Previously DNA sequences with
lots of ambiguous characters could be handled as proteins and search failed.
Fixing an issue with multiple placement of the same query. Adding a new option
“–output-discarded-queries” for one-by-one placement. Enabling placement to newly
added taxa now also for translated data. Speed-up for Exonerate search.
Adding ancestral sequence reconstruction for codon data when “–translate” is used.
A backup system for cases where BppAncestors fails.
Faster reference build (mainly for aa/codons). Less verbose output. Exonerate hits
trimmed to avoid overlapping hits.
By mistake, PAGAN pruned the files even if no option was selected. This is now fixed.
The program homepage has been re-written and PAGAN is now compatible with that. Several changes in placement heuristics and tree extension.
Cleaning up the code: many unused options removed. Xml output with nhx tags. Changed search order of external tools. Now using BppAncestors for ancestor reconstruction.
TID tags now considered also for ORFs. Fixing the double naming of ORFs. Changed a bad default value for branch-skip-probability (not fixing the bug yet!).
Consistent naming of nucleotide and protein output of ORF alignments. Improvements for alignment pruning. Option to trim extended alignment.
Functions for alignment pruning (to reduce large datasets). Anchoring coverage threshold for skipping the DP alignment altogether. Options ‘find-best-orf’ and ‘compare-reverse’ also with placement.
Improvements for pileup of 454 data. Removal of gaps in query sequences before alignment.
New option to create a distance tree using ‘bpp_dist’ package. Significant changes (and improvements) for pileup alignment of 454 reads.
Compiles on Darwin and Cygwin.
First integration of MAFFT and RAxML: PAGAN is now capable of computing a guide tree by first calling MAFFT and then RAxML.
Minor: code reorganised and recently created bugs fixed.
PAGAN can now anchor alignments with Exonerate or substring matching. As a result, significant speed ups in the alignment of long and/or very similar similar sequences.
Now properly functional multi-threading for standard multiple alignment: by default, PAGAN uses all the cores available. This can be changed with --threads #.
First attempt to introduce multi-threading for standard multiple alignment.
See VERSION_HISTORY in the source code