RAxML output and tree construction
0
0
Entering edit mode
3.3 years ago
yp19 ▴ 70

Hi all,

I've run RAxML and got all the output files as expected. My understanding is that RAxML_bestTree.results.tree contains the best scoring ML tree and RAxML_bipartitions.results.tree will produce the same tree but with support (bootstrap) values. When I plot these trees though, they are not identical (they are similar, but there are differences).

Best tree:

Bipartitions:

Does anyone know why that is happening?

Thanks!

phylogeny tree proteins • 3.0k views
0
Entering edit mode

What is RAxML command you used? Are you sure the differences are real, and not just some rotated branches?

0
Entering edit mode

Here is the command i'm using:

raxmlHPC-PTHREADS-SSE3 -T 8 -p 12345 -f a -s FcC_smatrix.phy -n results.tree -c 4 -m PROTGAMMAIJTTF -x 12345 -N 500


Yes i'm sure. I will try to upload the figures here to show them.

0
Entering edit mode

If I focus on one of the species (DDB) we can see it is located in different spots on the tree

best tree: https://i.ibb.co/Ny4S3Kn/besttree.png

bipartitions: https://i.ibb.co/Z1fCC34/bipartitions.png

0
Entering edit mode

Are your trees rooted or unrooted? At first glance, the topology seems identical, if the trees are unrooted.

edit: I updated your post to include the trees, hope you don't mind.

0
Entering edit mode

No problem. I don't think I specified rooted/not-rooted in my code (raxmlHPC-PTHREADS-SSE3 -T 8 -p 12345 -f a -s FcC_smatrix.phy -n results.tree -c 4 -m PROTGAMMAIJTTF -x 12345 -N 500 ), so it's most probably unrooted. Can you elaborate on how the topology is the same?

1
Entering edit mode

It is difficult to say for sure, as the tree tips don't have names. But it seems to me if you root both trees at the branch separating DDB and its closest species from the rest of the tree, the resulting trees with have identical topology. You can test this by using FigTree to root and compare the trees.

0
Entering edit mode

I will try that, thank you!!