Hello Sergei

I've been using Datamonkey to analyze my dataset and has helped me and gave me back some interesting results as well.

Something peculiar came out with one of the analysis where I used the same dataset for GA-branch and Branch-REL some of the branches have two different results.
I've attached a JPG of my results and hopefully it is uploaded and if you can see some of the nodes have two different colors (different omega values).
On the GA-branch result node2 and node3 although both should be one node are analyzed as two different nodes. Same happens in node 6 and 7.

I've used the same dataset for Branch-REL and something similar happens as well and I'am not sure what is happening.

Hopefully my question makes sense and I'am sorry for any confusion.

Sincerely Jae

---

Hi Jae,

This is interesting. Are you providing a tree to datamonkey or is it using a NJ/GARD tree? If the former, then I would appreciate seeing your alignment+tree file, because the "split" branch is likely a rerooting/unrooting artifact (a bug), which I would like to fix.

Sergei

Hi Sergei

So I've been supplying my own tree instead of the NJ tree.
Also I'am attaching the file I have used.

Thank you

Jae

---

Hi Jae,

Looks like your tree had an extra set of () forcing HyPhy to create a branch with one child, as in


Code (]
((((a,b),c)),d)
[/code):

How did you create it? The tree which does not have the phantom internal node is this [code] ((((((((((Dyak_bam,Dsan_bam),Dtes_bam),Dere_bam),(Dmel_bam,(Dsim_bam,Dsec_bam)))),(((Dtak_bam,Dbia_bam),Deug_bam))),Dfic_bam),(Dele_bam,Drho_bam)),Dkik_bam),(Dbip_bam,Dana_bam))

Sergei

Hi Sergei

Hmmm that is very strange. I've been building a custom tree using online newick tree builders and it must have got the additional () incorporated some how. But this makes more sense now.

I've asked another question in the bug section and maybe the additional () could have been the problem.

I'm sorry but I had one last unrelated question in that the NJ tree built by datamonkey have a very different tree when I use the program MEGA. I'm not sure how this could happen as I've used all possible nucleotide substitution models on MEGA and still did not get a tree that datamonkey builds. Did you have a suggestion to this?

Again thank you very much for the help.

Sincerely

Jae

Hi Jae,

Datamonkey uses the simple Tamura Nei 93 distance to build the phylogeny. Can you paste in the tree you get from MEGA for comparison?

Sergei

I've attached the MEGA tree built using TN93 and I'm beginning to think it might have been an issue with rooting the basal group.

Jae

---

Hi Jae,

I think this might be the same tree (up to sibling rearrangements) as what is reported by datamonkey.

Sergei

---

Yes it all makes sense now. Thank you for helping out again!

Sincerely

Jae