Node bars

[ghost]

Hi, how do we get node bars in the final dated tree? Than you, Best

[josephwb]

Hi Uzma.

Typically one bootstraps their alignment to generate 100+ bootstrap phylograms. Run treePL on each, then use some program (like treeannotator from the BEAST package) to generate a summary tree. Node bars can then be visualized in FigTree.

There is a question of whether to fix the topology during the bootstrapping. If the topology is fixed, then one is only accommodating uncertainty in edge lengths. But this make fossil placement simple. If topology is not fixed, then (of course!) topological uncertainty is accommodated. However, this can cause problems with placing fossil constraints. This all pertains to the situation when internal fossil constraints are used. If the only constraint is on the root, then there are no issues (and I would recommend not fixing topology).

HTH. :bowtie:

[ghost]

Thank you.. Unfortunately I'm not using fossil, but I didn't understand to fix topology, how could we fix topology?

[josephwb]

If you don't have fossils, don't worry about it. But any phylogenetics program has the option to estimate parameters (including edge lengths) on a fixed topology.

[ghost]

okay thank you Joseph. Best,

[stubbsrl]

The 100 bootstrap trees used to generate the BS values don't have branch lengths. Isn't that problematic for TreePL? What am I missing?

[rvosa]

I'm pretty sure that the idea is that the bars are computed around the node depths, hence the trees have to be ultrametric.

[stubbsrl]

Hmm okay. But how these bootstrap trees are rooted does matter, correct?

[rvosa]

Yes. At least, if I understand the general reasoning, you might do something like this:

• generate 100 bootstrapped replicates of the alignment
• using a fixed topology, estimate branch lengths for each bootstrap replicate
• run treePL on each replicate

The only variation you then have in the trees is that around the depths of the (non-fixed) nodes. Calculate node bars around those and you have bootstrapped node depths.

[stubbsrl]

Okay thanks - I think that is the step I was missing. I'm going to run raxml as part of a batch script for each of 100 bootstrap replicates of the alignment constrained my "correct" topology. Those output trees should have branch lengths and be rooted properly.

Thanks!!

[uzma-researcher]

Hello, I have question, for node bar or error bars, do I need to take bootstrap tree file a time in treePL in infile tree? Thank you!