converting from Mothur to Qiime

[Jean-Francois Martin]

dear all,
I did my OTU picking with mothur and converted the OTU table to biom with the command

make.biom(shared=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.pick.pick.an.unique_list.shared, constaxonomy=stability.trim.contigs.good.unique.good.filter.unique.precluster.abund.pick.pick.an.unique_list.0.03.cons.taxonomy)

the produced biom file works with a couple of Qiime (1.9.1) scripts as plot ones but I could not make it work with

core_diversity_analyses.py with an error : ValueError: Shape tuple is incompatible with data.

I could not find whether the problem comes from my biom or my mapping file.

I join the files that I think may be useful for anyone wanting to hep me. It may be something stupid but I could not find it after two days and did not find any related post on the forum yet.

I hope someone can help.

Thanks a lot,

Jef

[justink]

Odd. it appears to be specific to rewriting in hdf5 format. Looking into it.

[Jai Ram Rideout]

Hi Jef,

Can you please provide the following info?

1. The output from running: print_qiime_config.py -t

2. The exact core_diversity_analyses.py command you're running

Jai

[Husen Zhang]

I encountered the same problem and found a workaround: 

First convert the mothur-made biom to the classic tsv format, then back to json:

qiime

biom convert -i v1v3_subset_0.03.biom -o v1v3_subset_0.03.txt --to-tsv --header-key taxonomy

biom convert -i v1v3_subset_0.03.txt -o v1v3roundTrip.biom --to-json --process-obs-metadata taxonomy --table-type "OTU table"

summarize_taxa_through_plots.py -i v1v3roundTrip.biom -o test -m qiime_mapping_v1v3_subset.tsv -c Treatment

Husen

[Jean-Francois Martin]

First of all thank you both for your help, I greatly appreciate that.

@ Jai Ram Rideout : I attach the result of the config command. I use the virtual machine as provided at the Qiime web site.
the core_diversity command I use is : core_diversity_analyses.py -i file.biom -o ./output -m mapping.txt -e 8413 --nonphylogenetic_diversity
As you will see on the next line, the probleme is kind of solved but I am still curious to know where it comes from if you can find it I thnink it will be helpful to many people.

@ Husen Zhang : thanks a lot ! your workaround worked perfectly for me ! you save my day !

Thanks again,
Jef

[Jai Ram Rideout]

Thanks for the workaround Husen, and thanks for the details Jef! There are a couple of other workarounds in this thread, as well as details about the problem itself and why the workarounds are necessary.

Hope this helps,

Jai

[Jean-Francois Martin]

Thank you again. I am lucky that the Qiime community is very active!
Cheers