1
$\begingroup$

I apologize for this very basic question. With @Michael G.'s help, I have been learning how to use RAxML to build a phylogenetic tree (How to define an outgroup to build a robust amino acid tree) and after running my analysis I obtained several results files. I have both a RAxML_bestTree.result file and a RAxML_bootstrap.result file. How do I "combine" these files so that my tree shows the bootstrap values?

Any help will be greatly appreciated.

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ Brilliant, l'll post answer tomorrow, rather I'll post the code you need. Its really easy, all the code is within RAxML. Sorry for the delay $\endgroup$
    – M__
    Nov 19, 2019 at 20:28
  • 1
    $\begingroup$ Thanks a lot @Michael G! $\endgroup$
    – Leah
    Nov 21, 2019 at 1:30

2 Answers 2

Reset to default
1
$\begingroup$

Apologies for the delay. The calculation is a majority consensus phylogeny of a maximum likelihood bootstrap, which is superimposed into a maximum likelihood phylogram. That's the tech. speak over ... the command is,

raxmlHPC -f b -m PROTGAMMAILG -n output_bootstrap.tre -t RAxML_bestTree* -z RAxML_bootstrap.result
  • RAxML_bestTree* = your maximum likelihood tree without bootstraps (file is probably 1 line long)
  • RAxML_bootstrap.result = your complete list of bootstraps (file is 500 lines long [or the number of bootstraps you set])
  • output_bootstap.tre = final result tree, i.e. majority consensus. This file is one line long but will have with all bootstraps attached to each branch.

The calculation will take less than one minute. The code is C and very slick. Note the "RAxML_bestTree*" is probably better to write the file name in full, not truncated via the glob (*). BTW you must specify the model.

Visualising the tree You use a Mac ... so the cool tree program is figtree . Make sure when you load it you click the branch option and set it to see your bootstraps, otherwise it will not visualise them by default. You might need to re-root the tree, from memory your command specified an outgroup. It might not be the right one though and figtree will easily re-root it for you. Bootstrap values above 75% are considered robust. If you get stuck, can chat but I'm not around until this time next week (bereavement leave and loads of meetings).

Improve this answer
$\endgroup$
5
  • 1
    $\begingroup$ Thank you very much for the detailed answer @Michael G. I will give it a try and see how it goes. If I get stuck I will let you know to see if we can set up a chat next week. Again, I appreciate your taking the time to help me out, especially at this time. I am sorry for your loss. $\endgroup$
    – Leah
    Nov 22, 2019 at 1:53
  • $\begingroup$ Hi @leah just let me know, if you have a inquiry about the above please post here, I can answer quickly. If it's an involved explanation chat room is better (end of next week) $\endgroup$
    – M__
    Nov 22, 2019 at 16:36
  • 1
    $\begingroup$ I'm sorry for the delay @Michael G. I was finally able to run the command (it is very fast indeed) and I got three output files: RAxML_bipartitions.output_bootstrap.tre, RAxML_bipartitionsBranchLabels.output_bootstrap.tre and RAxML_info.output_bootstrap.tre. I opened the RAxML_bipartitions.output_bootstrap.tre file in FigTree and I got a tree with bootstrap values. However, I couldn't open the RAxML_bipartitionsBranchLabels.output_bootstrap.tre file: Error reading file: Number format error: For input string: "0.31..21[98]". What is the difference between these files? Thank you for your help! $\endgroup$
    – Leah
    Dec 4, 2019 at 2:10
  • $\begingroup$ Hi Leah in summary you just need RAxML_bipartitions.output_bootstrap.tre. If you want a detailed answer could you ask this as a separate question please? Alternatively we can chat, but it is the sort of thing that is better as a question, because I'll need to think about how to present the information and it might be useful to others. Its a bit more complicated and theoretical than command line trouble shooting. The question title is something like "interpreting quantitive outputs from maximum likelihood phylogenetic trees" and the information for the question is what you have just asked above $\endgroup$
    – M__
    Dec 5, 2019 at 0:35
  • 1
    $\begingroup$ Thanks @Michael G! The question is posted here: bioinformatics.stackexchange.com/questions/10948/… $\endgroup$
    – Leah
    Dec 6, 2019 at 15:07
2
$\begingroup$

Depending on the search parameter settings RAxML produces several output files among which you normally find:

RAxML_bipartitions.result: If you used the ­-f b option, this file will contain the input tree with confidence values from 0 to 100 drawn on its nodes! It is also printed when -f a ­x have been specified, at the end of the analysis the program will draw the BS support values on the best tree found during the ML search.

RAxML_bipartitionsBranchLabels.result: Contains the same information as the file above, but support values are correctly displayed as Newick branch labels and not node labels! Support values always refer to branches/splits of trees and never to nodes of the tree. Note that, some tree viewers have problems displaying branch labels, they are however part of the standard Newick format.

from https://cme.h-its.org/exelixis/resource/download/NewManual.pdf

If you don't find the files in your output directory you did not specify the above parameters. If a rerun is not feasible you could you use the multinewick file RAxML_bootstrap.result with raxml or as input for the program splitstree to generate a consenus tree with different rules (majority rule, strict, etc.). However, these options don't guarantee to generate the topology of the maximum likelihood tree.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Thank you very much for your help @bajah, I really appreciate it! $\endgroup$
    – Leah
    Nov 21, 2019 at 1:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.