Error when consenus tree for incongruence calculations

[Zachary Harris]

Hi All,

I have 193 genes trees that have been bootstrapped using RAxML. I passed all of these bootstrapped trees to Astral and generated a consensus tree:

(Leaf_1,(Leaf_2,((Leaf_3,Leaf_4)1:0.669538512836948,((Leaf_5,(Leaf_6,Leaf_7)1:0.15474512669113813)1:0.0580634945409515,:Leaf_8)1:0.08439709160748539)1:0.20546201163939506));

I am now interested in calculating incongruency metrics on the 193 individual gene trees when compared to the consensus. However, when I try this, I get the following error:

"ERROR: Expecting ':' in tree; found: character ','"

I also tried generating a consensus tree using RAxML, but it gave the same error. Does anyone have any sight as to the cause/ solution of this problem?

For explicitness, the command that is failing is as follows
raxmlHPC-SSE3 -f i -m GTRCAT -t astral.tree -z RAxML_bootstrap.m0.T1 -n T2

Thanks,
ZH

[Alexandros Stamatakis]

Hi Zachary,

RAxML expects a fully bifurcating tree as input, but there also seems to
be a problem with the tree format here:

",:Leaf_8)"

is not valid Newick.

You should try running RaxML -f i with a species tree as input.

Alternatively, you can use a new method for calculating TC/IC-like
measures that should go on-line on biorxiv and whose code should also
become available today. I shall post this on here later-on.

Alexis

> --
> You received this message because you are subscribed to the Google
> Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to raxml+un...@googlegroups.com
> <mailto:raxml+un...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology

www.exelixis-lab.org

[Alexandros Stamatakis]

And here is the new method I mentioned earlier today:

http://www.biorxiv.org/content/early/2017/07/27/168526

Alexis

[Heiko]

Hi Zachary, hi Alexis,

> (Leaf_1,(Leaf_2,((Leaf_3,Leaf_4)1:0.669538512836948,((Leaf_5,(Leaf_6,Leaf_7)1:0.15474512669113813)1:0.0580634945409515,:Leaf_8)1:0.08439709160748539)1:0.20546201163939506));
>
> I am now interested in calculating incongruency metrics on the 193
> individual gene trees when compared to the consensus. However, when I
> try this, I get the following error:
>
> "ERROR: Expecting ':' in tree; found: character ','" 

 
[...]

> be a problem with the tree format here:
> ",:Leaf_8)"
> is not valid Newick.
 

I spotted that one as well, which is definitely an error, but it actually does not fit the error message.
The message says that a comma (',') is found instead of an expected colon (':').
However, at Leaf_8 we find a colon (':') which does not belong there (thus, not expected).

But there is another issue where we find commas (',') instead of colons (':') - All leaf branches do not have any branch-lengths which would follow the leaf name separated by a colon. Instead after Leaf_1, Leaf_2, Leaf_3,Leaf_5 and Leaf_6 there is a comma instead (plus more that have closing brackets instead.
Would this also be a candidate for causing that error?

Just, out of interest was that tree created by a program or manually? The latter would explain the error at Leaf_8 which I would not expect from a software.

Best regards,
Heiko

[Zachary Harris]

Hi All, 

The mistake t about Leaf_8 was definitely a mistake on my part. That isn't in the tree that I pass to RAxML. The "," that it seems to have a problem with follows Leaf_2. 

I do agree that is it is presumably a problem with branch lengths. The tree was generated using Astral. 

Zach

[Alexandros Stamatakis]

Hi Zach,

Without the original Astral output tree we can't really help you.

Nontheless, I'd encourage you to use the new quartet-based method that
we made available yesterday as it seems to perform better for your
purpose anyway.

All the best,

Alexis