Low Bootstrap values from online RAXML

[Susie Ramsey]

I did a bootstrap analysis of only about 10 bootstraps on the online RAXML (because I can't download it) and it gave me the files associated with my tree. I have opened up reult.raxml.support, because in the file it said that's where the bootstraps were located, in FigTree and it gave me branch annotations that are between 0-2 and are long decimals. I don't understand why these numbers would be so low?

Is it because of the low number of repeats but i thought it would have to be only number of 10-100 if there was only 10. I saw it could be because it is mtDNA and they're closely related indivuduals of the same species so could cause low numbers?

Thanks

[Grimm]

Hi Susiera,

this sounds like you did 10 tree inferences but not really a bootstrapping analysis. You are right: If you would have only 10 bootstrap pseudoreplicates, the minimum support would indeed need to be 10 (the split occurred in a single of the 10 replicate trees) unless there's a bug somewhere (standard setting is usually 100, and with tricky data, one can make use of the bootstop criterion to make sure the number of bootstrap replicates is sufficient).

Another issue maybe with FigTree itself, what you select to display.

Regarding your question: Absolutely, if you use mtDNA of closely related individuals and classic barcode marker (like cox1, cytB etc), the signal in your matrix may be too faint for ML tree inference and bootstrapping. If your data in addition includes many near-identical tips, those so-called flat subtrees will have no branches with high support. If you work below the species level, the stochastic population dynamic processes will inflict a lot of terminal noise, which may obscure the phylogenetic structuring in your data set further.

PS Some essentials to get more concrete help:

• Give the RAxML version you (the online server) used
• Post the command you used (if not entirely handled by the online server)
  
• Post the apparently odd result file(s)
  

Cheers, Guido

[Alexandros Stamatakis]

> I did a bootstrap analysis of only about 10 bootstraps on the online
> RAXML (because I can't download it)

You can download RAxML-NG here:

https://github.com/amkozlov/raxml-ng

> and it gave me the files associated
> with my tree. I have opened up reult.raxml.support, because in the file
> it said that's where the bootstraps were located, in FigTree and it gave
> me branch annotations that are between 0-2 and are long decimals. I
> don't understand why these numbers would be so low?

Could you send us the file please?

> Is it because of the low number of repeats but i thought it would have
> to be only number of 10-100 if there was only 10.

No it is not because of the low number of bootstrap replicates.

> I saw it could be
> because it is mtDNA and they're closely related indivuduals of the same
> species so could cause low numbers?

That is a probable cuas (i.e., lack of signal), you can actually test
for this a bit using the bootstrap convergence criterion that will do as
many bootstrap replicates as necessary until the support values
stabilize, if it stops after a large number of replicates this is always
an indication that there is not much signal in your data.

Alexis

>
> Thanks
>
> --
> 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>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/raxml/aa76d2e9-446a-41b1-a022-5ed60ce4e1e7n%40googlegroups.com
> <https://groups.google.com/d/msgid/raxml/aa76d2e9-446a-41b1-a022-5ed60ce4e1e7n%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
Affiliated Scientist, Evolutionary Genetics and Paleogenomics (EGP) lab,
Institute of Molecular Biology and Biotechnology, Foundation for
Research and Technology Hellas

www.exelixis-lab.org