convert nexus to newick

[monkey79]

Hi there,

I need to convert my trees from nexus format to newick format. Is that
possible with ete2?

thanks!

[Daniel Pérez Martín]

Having found nothing in ETE2 docs, you might want to code a conversion utility using both ETE2 and a python library for nexus such as python-nexus:

http://simon.net.nz/articles/python-nexus-a-generic-nexus-nex-trees-reader-for-python/

Regards,

Daniel.

2012/3/29 monkey79 <barbara...@gmail.com>

[Jaime Huerta Cepas]

There is no built-in nexus parser in ETE since Nexus (in the case of trees) seems to be a mere wrapper of the newick format (http://www.bioperl.org/wiki/Nexus_tree_format).

You can use any parser such as the one suggested by Daniel to extract the newick and then load it using ETE. Alternatively, you can just parse the Nexus file to extract lines starting with "tree". This is a simple example using regular expressions:

import re
from ete2 import Tree

nexus = """
#NEXUS
begin trees;
tree 'name' =(1:0.212481,8:0.297838,(9:0.222729,((6:0.201563,7:0.194547):0.282035,(4:1.146091,(3:1.008881,
(10:0.384105,(2:0.235682,5:0.353432):0.323680):0.103875):0.413540):0.254687):0.095341):0.079254):0.000000;

tree 'tree2' =(A:0.212481,B:0.297838,(C:0.222729,((D:0.201563,7:0.194547):0.282035,(4:1.146091,(3:1.008881,
(10:0.384105,(2:0.235682,5:0.353432):0.323680):0.103875):0.413540):0.254687):0.095341):0.079254):0.000000;


end;
"""

match  = re.search("begin trees;.+end;", nexus, re.MULTILINE | re.DOTALL)
if match:
    tree_block =  match.group()
    for match in re.finditer("\s*tree\s+([^=]+)=([^;]+;)", tree_block, re.MULTILINE | re.DOTALL):
        tree_name = match.groups()[0].strip()
        nw = match.groups()[1]
        t = Tree(nw)
        print t
   
If any of you has something more elaborated, I will be happy to include it in ETE.

Jaime

2012/3/29 Daniel Pérez Martín <dpe...@gmail.com>

[monkey79]

Hey Jaime,

here is an example of what my nexus tree files likes. it is the result of a bayesian analysis. I have got about 1700 trees and I am trying to figure out how resolve into a newick format in order to use ete2. Any other suggestion?

thanks

#NEXUS

[ID: 4199812097]

begin taxa;

dimensions ntax=5;

taxlabels

ENSTRUG00000000417

T18D3.7

ENSGACG00000013730

ENSLACG00000012925

ENSLACG00000003805

;

end;

begin trees;

translate

1 ENSTRUG00000000417,

2 T18D3.7,

3 ENSGACG00000013730,

4 ENSLACG00000012925,

5 ENSLACG00000003805

;

   tree con_all_compat = [&U] (ENSTRUG00000000417[&prob=1.000000000000000e+00,prob_stddev=0.000000000000000e+00,prob_range={1.000000000000000e+00,1.000000000000000e+00},prob(percent)="100",prob+-sd="100+-0"]:1.418146859482858e-01[&length_mean=1.453808394263178e-01,length_median=1.418146859482858e-01,length_95%HPD={4.394102925423538e-02,2.398851568109936e-01}],ENSGACG00000013730[&prob=1.000000000000000e+00,prob_stddev=0.000000000000000e+00,prob_range={1.000000000000000e+00,1.000000000000000e+00},prob(percent)="100",prob+-sd="100+-0"]:7.103231066152176e-01[&length_mean=7.193836401649210e-01,length_median=7.103231066152176e-01,length_95%HPD={5.013425489412863e-01,9.651804467297473e-01}],((T18D3.7[&prob=1.000000000000000e+00,prob_stddev=0.000000000000000e+00,prob_range={1.000000000000000e+00,1.000000000000000e+00},prob(percent)="100",prob+-sd="100+-0"]:9.105574486435167e-01[&length_mean=9.257404980325880e-01,length_median=9.105574486435167e-01,length_95%HPD={5.933291691904896e-01,1.277730934852635e+00}],ENSLACG00000012925[&prob=1.000000000000000e+00,prob_stddev=0.000000000000000e+00,prob_range={1.000000000000000e+00,1.000000000000000e+00},prob(percent)="100",prob+-sd="100+-0"]:5.505290596097889e-01[&length_mean=5.614284096694910e-01,length_median=5.505290596097889e-01,length_95%HPD={3.098368159543265e-01,8.231827588050069e-01}])[&prob=9.504132231404958e-01,prob_stddev=2.262138462873564e-03,prob_range={9.488136496934151e-01,9.520127965875766e-01},prob(percent)="95",prob+-sd="95+-0"]:2.232700123422474e-01[&length_mean=2.306815188221150e-01,length_median=2.232700123422474e-01,length_95%HPD={3.964294057691813e-02,4.360923788012644e-01}],ENSLACG00000003805[&prob=1.000000000000000e+00,prob_stddev=0.000000000000000e+00,prob_range={1.000000000000000e+00,1.000000000000000e+00},prob(percent)="100",prob+-sd="100+-0"]:8.688248742304949e-02[&length_mean=9.037320533212091e-02,length_median=8.688248742304949e-02,length_95%HPD={6.775268382584457e-03,1.727273852136540e-01}])[&prob=4.964009597440682e-01,prob_stddev=5.655346157183772e-03,prob_range={4.924020261263663e-01,5.003998933617702e-01},prob(percent)="50",prob+-sd="50+-1"]:6.385319910354678e-02[&length_mean=7.043041817985121e-02,length_median=6.385319910354678e-02,length_95%HPD={1.049211220468756e-05,1.546138043511682e-01}])[&prob=1.000000000000000e+00,prob_stddev=0.000000000000000e+00,prob_range={1.000000000000000e+00,1.000000000000000e+00},prob(percent)="100",prob+-sd="100+-0"][&length_mean=0.000000000000000e+00,length_median=0.000000000000000e+00,length_95%HPD={0.000000000000000e+00,0.000000000000000e+00}];

end;

[Jaime Huerta Cepas]

It doesn't look like a standard newick format at all :). You will probably need some regular-expression-magic to adapt the format to the extended newick and conserve the extra node info. 

If you just want the topology of the tree, the attached script seems to work with your example. 

Which program is generating the newick tree? 

[monkey79]

it is Mr.bayes

thanks Jaime

[banta....@gmail.com]

Hello,

I ran into this post trying to solve the same problem. I figured it out! You need to use a program called TreeGraph 2 to perform the conversion. It's the only program (besides FigTree, which won't output the node labels) that can read Mr. Bayes output files properly. I made a tutorial detailing how to do the conversion here:

https://www.youtube.com/watch?v=nHeH_oPGIIA

it is Mr.bayes

Daniel.

2012/3/29 monkey79 <barbar...@gmail.com>