Generate a tree (.tre) file through closed reference OTU picking

[Paz Aranega]

Hi,

I am analysing V4 16S Myseq data. I have done closed reference OTU picking after passing my samples through split_libraries_fastq.py both with the default parameters and with the SILVA database. It seems to have worked and I get a OTU table. However to perform further analysis using either the core_diversity_analysis.py script or any other scripts to calculate alfa or beta diversity I seem to need trees but I haven't got any file with the .tre extension. Is there any extra step I need to do to generate trees if I am doing closed-reference OTU picking?

Thanks,

Paz

[Yoshiki Vázquez Baeza]

Hello Paz,

If you are using the default reference database, then you should be able to use the Greengenes 97% tree, it should have been included as part of the output folder. Alternatively, you can download the reference files here ftp://greengenes.microbio.me/greengenes_release/gg_13_5/gg_13_8_otus.tar.gz

As for SILVA, I don't believe the authors provide a tree file due to the technical challenges that would imply, so I don't think you would be able to use a tree. However not all is lost, remember there are other non-phylogenetic metrics implemented in QIIME, so you would be able to use those without a problem.

Thanks!

Yoshiki.

[Qi Yang]

Hi Paz,

I have the same problem as yours. Did you solve the problem? Could you please let me know how i can generate a tree after the command pick_closed_reference_otus.py? So far, i only have a biom file after this command.

Thank you very much.

Qi

[Jamie Morton]

Hi Qi,

You shouldn't have to generate the tree.  You should just be able to use the reference 97% Greengenes tree: ftp://greengenes.microbio.me/greengenes_release/gg_13_8_otus/trees/97_otus.tree

The reason why you'd want to use a reference tree is because the reference tree was built using the entire 16S gene. 

If you were to generate the tree yourself, you'd be building the tree just based on the V4 region - which is just a fragment of the 16S gene.

Because of this, the greengenes tree most likely going to be more robust than building the tree based on short V4 based reads.

Concerning SILVA, I'd check out the qiime compatible files here.  It looks like they also have reference trees stored in there.

Hope that helps!

Best,

Jamie

[Qi Yang]

Hi Jamie,

Thank you very much for your reply.

May I confirm whether i got correct understanding? I only need download the tree file from the database as a guidance to do the alpha & beta diversity analysis instead of a tree generated from my input data? Because i would like to analyse the alpha using the PD_whole_tree metric and beta diversity using UniFrac metric. Is that ok to use the reference tree to do it?

Normally, i obtained the tree after the otu picking (open_reference or de_novo). So i though for 'closed_reference otu picking', the tree file also can be created automatically. But i only got a BIOM file.

Thank you,

Qi

[Jamie Morton]

open_reference and de_novo picking are a little different - you are trying to estimate OTUs on the fly.  In those scenarios you would use the tree generated by those methods.

With closed_reference, you are picking OTUs against a reference database.  Since that reference database already has a high quality tree bundled with it, you don't need to rebuild the tree.

So yeah, when you are doing closed ref picking, you can just download that tree, and input that tree for your alpha and beta diversity analyses.

Best,

Jamie

[Qi Yang]

Hi Jamie, 

Thank you very much for your reply. I will have a try. 

Qi