possible to make a tree from biom file?

[Ming Liao]

Hello, there

I am trying to make alpha-diversity analysis for metagenomic data. Currently, I can get a biome file for each metagenomic sample from metaphlan2, then all the samples have been merged into "all_sample.biom". 

Until now, I can use alpha_diversity.py  to calculate alpha diversity on each sample in "all_sample.biom". However,  when I chose PD_whole tree among the alpha diversity metrics, a tree file is required. How can I get a tree file in this case? 

Thanks

Ming

[jonsan]

Hi Ming,

As far as I know there's not a simple way to do this. MetaPhlan uses a taxonomic rather than a phylogenetic notion of bacterial diversity, so as far as I know there's not an easily accessible phylogenetic tree precomputed for the 'OTUs' (really microbial taxonomic labels) used as identifiers in the biom table. 

One option would be to use the taxonomy tree implied by the hierarchical ordering of taxonomy labels encoded within the identifiers themselves. It looks like you might be able to get this via the GraphLan output functionality script provided with MetaPlan. This wouldn't be a true phylogenetic tree, but it would give you at least notion of relatedness that might be informative. 

As a final alternative, you could choose a single genome from each of the terminal categories in the version of the MetaPhlan database you used and build your own phylogenetic tree de novo. 

Cheers,

-jon