Galaxy Taxonomy Output To Newick Tree

Question

[Zach Powers]

Hi Biostars,

I would like to take the output of Galaxy's Metagenomic Analyses > Fetch Taxonomic representation and make a Taxonomic tree that contains abundance information. For display, I could then experiment with the the many excellent visualization softwares out there ( i am particualry interested in the ETE2 library) . This question is very similar to a previous question where the author was reccommended to use iTOL for a similar purpose. I am wondering, however, if it possible to generate an abundance-information-containing Newick tree directly from the Galaxy Metagenomic Analyses output - ideally to provide a flexible replacement for the current taxonomic abundance-visualization tool. The values in the Newick tree could actually come from the

The data output is a table with a few identifier collumns and then the kindom/phlyum/class.... Here are a few lines (sorry that it wraps):

contig00428    562    root    Bacteria    n    n    n    Proteobacteria    n    n    Gammaproteobacteria    n    n    Enterobacteriales    n    n    Enterobacteriaceae    n    n    n    Escherichia    n    Escherichia coli    n    42469
contig00073    562    root    Bacteria    n    n    n    Proteobacteria    n    n    Gammaproteobacteria    n    n    Enterobacteriales    n    n    Enterobacteriaceae    n    n    n    Escherichia    n    Escherichia coli    n    42945
contig00672    562    root    Bacteria    n    n    n    Proteobacteria    n    n    Gammaproteobacteria    n    n    Enterobacteriales    n    n    Enterobacteriaceae    n    n    n    Escherichia    n    Escherichia coli    n    146126
contig00143    562    root    Bacteria    n    n    n    Proteobacteria    n    n    Gammaproteobacteria    n    n    Enterobacteriales    n    n    Enterobacteriaceae    n    n    n    Escherichia    n    Escherichia coli    n    287840
contig01215    562    root    Bacteria    n    n    n    Proteobacteria    n    n    Gammaproteobacteria    n    n    Enterobacteriales    n    n    Enterobacteriaceae    n    n    n    Escherichia    n    Escherichia coli    n    290448

Alternatively the 'Summarize Taxonomy tool' provides the following:

superkingdom    Eukaryota    9
superkingdom    Viruses    1
kingdom    Fungi    2
kingdom    Metazoa    5
kingdom    Viridiplantae    1
subkingdom    Dikarya    2
superphylum    Bacteroidetes/Chlorobi group    22
superphylum    Chlamydiae/Verrucomicrobia group    6
superphylum    Fibrobacteres/Acidobacteria group    8
phylum    Acidobacteria    8
phylum    Actinobacteria    9
phylum    Apicomplexa    1

any ideas would be appreciated, zach cp

[jhc]

Zach, how big (# of terminal nodes) are your trees?

[Zach Powers]

jhc, the tree varies depending on the experiment but a few thousand is typical. I actually came across your Taxonomy Lookup and have been playing with it. It is similar to Pierre's but outputs the newick tree which I can then use with the ETE2 library for custom layout. (which is awesome - thanks!).

Answer 1

[Pierre Lindenbaum]

The following java code: https://gist.github.com/2787783 takes two parameters:

• a directory for the dump of NCBI taxonomy

• a file containing a list of taxonomy-id

  javac -Biostar45691.java java Biostar45691NCBITAXONOMY taxonsidlist.txt > result.gexf

it then produces a XML-based Graph (Gexf) .

<graphml xmlns="&lt;a href=" http:="" graphml.graphdrawing.org="" xmlns"="" rel="nofollow">http://graphml.graphdrawing.org/xmlns" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://graphml.graphdrawing.org /xmlns <http: graphml.graphdrawing.org="" xmlns="" 1.0="" graphml.xsd"&gt;&gt;="" &lt;key="" id="taxname" for="node" attr.name="Taxon.Name" attr.type="string" &gt;="" &lt;key="" id="count" for="node" attr.name="Count" attr.type="int" &gt;="" &lt;key="" id="countleaf" for="node" attr.name="CountLeaf" attr.type="int" &gt;="" &lt;graph="" edgedefault="directed" &gt;="" &lt;node="" id="28216" &gt;="" &lt;data="" key="taxname" &gt;Betaproteobacteria&lt;="" data&gt;="" &lt;data="" key="count" &gt;1&lt;="" data&gt;="" &lt;data="" key="countleaf" &gt;0&lt;="" data&gt;="" &lt;="" node&gt;="" &lt;node="" id="42256" &gt;="" &lt;data="" key="taxname" &gt;Rubrobacter="" radiotolerans&lt;="" data&gt;="" &lt;data="" key="count" &gt;1&lt;="" data&gt;="" &lt;data="" key="countleaf" &gt;1&lt;="" data&gt;="" &lt;="" node&gt;="" &lt;node="" id="2748" &gt;="" &lt;data="" key="taxname" &gt;Carnobacterium="" divergens&lt;="" data&gt;="" &lt;data="" key="count" &gt;1&lt;="" data&gt;="" &lt;data="" key="countleaf" &gt;1&lt;="" data&gt;="" &lt;="" node&gt;="" &lt;node="" id="2" &gt;(...)<="" p="">

You can open this kind of graph using gephi and visualize the usage of each taxon.

In the example below, I've displayed a few ~100 nucleotide having "16S RNA" in their title. You can use the number of times a taxon and his parents (lower) were found or only the count of each taxon from your list (upper)

[Zach Powers]

Pierre, thank you. This is basically what I am looking for except that I am unfamiliar with the gexf format. Is it possible to control tree shape/appearance? -- (I don't see anything like a phylogenetic tree in the Gephi documentation.)

[Pierre Lindenbaum]

gexf is a simple XML-based format describing a graph (nodes+edges) and used by gephi. Once opened in gephi, you can customize the graph (colors, layout, etc...) http://gexf.net/format/basic.html