'Backwards' branches in some trees

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Run a MrBayes mcmc including a taxon constraint that contradicts a grouping 
found in the maximum likelihood tree
2. Open the resulting .run1.t file
3. Draw root canal
4. Taxa appear in a position that violates the constraint, and the branch 
leading to some of the constrained taxa goes from right to left (not left to 
right)

What is the expected output? What do you see instead?
1. Should groups that violate the constraint appear in the root canal?
2. Should branches go from right to left, and how do I interpret that?

What version of the product are you using? On what operating system?
2.01 on Windows XP

Please provide any additional information below.

Let me know if you need more information...

Original issue reported on code.google.com by MartinS on 29 Feb 2012 at 4:46

[GoogleCodeExporter]

Hi Martin,

Sorry for the slow response, I did not realise issues are reported here so 
forgot to look for it.

The root canal tree has a topology that is selected from the set of trees in 
the tree set and has the highest probability of clades in the set. The height 
of the nodes in the root canal tree are determined by the average of the height 
of the clade. If some of the clades in the tree have low support it can happen 
that the clade's height is below that of on of its children. This is a well 
known issue with this kind of summary tree, and I would not consider it a 
defect.

So, backward branches are a sign that clades in the root canal tree are not 
well supported by the data.

Original comment by higgs...@gmail.com on 5 Jul 2012 at 3:51

[GoogleCodeExporter]

Thanks for the explanation!
- Martin

Original comment by MartinS on 5 Jul 2012 at 8:41