Let’s consider the following phyloXML sample:
<phyloxml
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns="http://www.phyloxml.org"
xsi:schemaLocation="http://www.phyloxml.org http://www.phyloxml.org/1.10/phyloxml.xsd"
>
<phylogeny rooted="true">
<name>Alcohol dehydrogenases</name>
<description>contains examples of commonly used elements</description>
<clade>
<events>
<speciations>1</speciations>
</events>
<clade>
<taxonomy>
<id provider="ncbi">6645</id>
<scientific_name>Octopus vulgaris</scientific_name>
</taxonomy>
<sequence>
<accession source="UniProtKB">P81431</accession>
<name>Alcohol dehydrogenase class-3</name>
</sequence>
</clade>
<clade>
<confidence type="bootstrap">100</confidence>
<events>
<speciations>1</speciations>
</events>
<clade>
<taxonomy>
<id provider="ncbi">1423</id>
<scientific_name>Bacillus subtilis</scientific_name>
</taxonomy>
<sequence>
<accession source="UniProtKB">P71017</accession>
<name>Alcohol dehydrogenase</name>
</sequence>
</clade>
<clade>
<taxonomy>
<id provider="ncbi">562</id>
<scientific_name>Escherichia coli</scientific_name>
</taxonomy>
<sequence>
<accession source="UniProtKB">Q46856</accession>
<name>Alcohol dehydrogenase</name>
</sequence>
</clade>
</clade>
</clade>
</phylogeny>
</phyloxml>
It can be mapped to the following Jevko:
which contains identical information, is not less human-readable, and at the same time is much more efficient to process, because it is simpler and smaller (less than 79% of the size of the original XML).