Read this first: download links are at the bottom.
Updates To Clann:
20th May 2004:
Update implementing the Readline library for Linux and Mac OSX versions of the software and addressing some minor bugs regarding the interface.
16th November 2004
Released Readline enabled and disabled versions of Clann for Mac OSx.
Red Hat Linux version of Clann does not require a readline installation.
- Random sampling of treespace to better choose a starting tree for heuristic searches.
- Settings displayed to screen before each command is carried out.
- Neighbor-joining tree of MRP matrix (which can be used as a starting tree for heuristic searches).
- Bug fixed: maxswaps option not working in SPR heuristic searches
- Bug fixed: Clann crashing with large datasets.
- Clann Manual updated to cover new features and some features that were left out of earlier versions.
30th September 2005 (release of Clann 3.0)
- Fixed Several bugs
- New Criterion added: Average Consensus. Missing Data estimated using either an ultrametric or 4 point condition estimate.
- NJ tree can be created using the command 'nj'. This is a neighbor-joining tree of an average consensus distance matrix of the data.
- Neighbor-joining tree can be used as a starting tree for heuristic searches.
- 'Showtrees' command added. This allows the user to view selections of the sourcetrees as ascii trees on the screen. The user can select the trees to be displayed according to their size, taxa compliment, score against the best supertree or according to the name of the trees. The selection can be saved to file.
- Specific weights can be assigned to individual source trees.
- 'Excudetrees' and 'Includetrees' commands added. This allows the user to either exclude or include trees from the analysis according to the size of the tree, the taxa compliment, the score of the tree against a predefined supertree or according to the name given to the tree.
- 'Deletetaxa' command added. This allows the user to select a taxon for deletion from the dataset. This has the effect of pruning the taxa from each of the source trees in which it appears. Source trees that have fewer than 4 taxa after any taxa are deleted are removed from the dataset.
- 'Generatetrees' command added. This allows the user to generate random supertrees and assess them against the source tree data (or against randomised of idealised versions of the source trees). The results from the generation of supertrees are displayed as a histogram. Statistics about the distribution are also calculated. This command also allows the user to create (and save to file) randomised or idealised version of their data.
- 'Consensus' command added. This allows Clann to carry out a consensus analysis of all the universally distributed trees (those trees that have all the taxa) in the source tree dataset. The user can choose from doing a majority-rule, or strict consensus. The results are displayed to screen in a graphical ascii format. This means that clann can now be used for the summary of bootstrap proportions from phylogenetic analyses.
- Clann now also calculates a consensus tree from any bootstrap analyses it carries out.
- Robinson-foulds distances can now be calculated from the sourcetrees dataset using the command 'rfdists'. For each comparison between two source trees, taxa are pruned from both trees until only the taxa that are shared by both trees remain. The results of these calculations are saved to file. The user has the choice of matrix or vector format.
- Clann now reads nexus format files. The file may contain translation tables or not and this also allows the inclusion of a 'clann block'. This is a set of commands that are to be carried out on the data in the file. Clann will ignore any data blocks that are not relevant to it.
- When a dataset is executed (loaded into memory) Clann will now also display a summary of the distribution of sizes of trees.
- Clann will now "catch" ctrl-c signals so that a search of tree-space may be aborted early. Clann will display the best supertree found before the ctrl-c signal was sent.
8th November 2005 (release of Clann 3.0.1)
- A modification of the seed choice for the random number generator was made to make sure that if several Clanns are started up at once, they will all have different seeds
- Fixed a bug in the robinson foulds distances. This over estimated the distances if the input trees given were rooted. The bug has been fixed.
- Clann was reported to crash with some large datasets, this has been rectified.
- A problem reading nexus files has been fixed.
- Some problems with aspects of the "include" and exclude" commands have been fixed.
- Some minor cosmetic changes were made to the interface.
27th November 2005 (release of Clann 3.0.2)
- A Bug which caused Clann to crash with some datasets have been resolved. This bug was introduced in the version 3.0.1
- Changes to the SPR heuristic when using the DFIT criterion has increased its speed. Inital tests indicate the speed increase can be of the region of 30% to 60% depending on the dataset used
Contact Chris at this email address
Download the required version here: