Hi Karl,

Yes it is possible, but depending on your study design it might be beneficial to to start with a hypothesis e.g. "Is nitrogen fixation altered?"  And then focus only on the involved genes in order to reduce correction for multiple testing. 

You could give it a try at the Huttenhower's Galaxy server. You might also have a look at tax4fun which also predicts functions from 16S rRNA gene fragments, but allows the use of Silva instead of the outdated GreenGenes database 

Anitha, do you know why a fasta or qual file won't upload to the server after registration...?

Hi Karl,

Regarding the upload, I'm assuming that your files are bigger than 500MB or the user quota.

For PICRUSt you only need your closed reference biom table. 

have a look at 

http://picrust.github.io/picrust/tutorials/otu_picking.html

Thank you Mina! 

I will try that! Have you done much work in picrust?

Do you run it in a Linux flavour or favour the user interface environment such as windows?

:)

Hi Karl,

I ran it a couple times on mac and on a galaxy server. Unfortunately it's not available on Windows, but you could still run it under virtualbox on windows (such as the qiime virtualbox image). If you prefer to use mothur instead of qiime's closed reference, you could follow their de novo PICRUSt SOP.

https://github.com/rprops/PICRUSt_from_mothur/blob/master/README.md

Best Mina

Hello Mina,

So let me get this straight. If I connect to https://huttenhower.sph.harvard.edu/galaxy/

through Windows and upload a OTU_Table.biom file it won't give any out put files?

I guess I am just wondering because, I succesfully uploaded some biom files but they wouldn't run through the hub server. I was thinking it was a formatting error on my end and I may need to run the biom file through the QIIME pipeline to get it into the right format, but now you said that it isn't available on Windows so I am a little confused.

I appreciate  your comments nonetheless as they are very helpful!

Karl

Hi Karl,

did you make sure to  select "picrust" while uploading? It needs to have this data type to show up in normalize by copy numbers. You could also switch the data type after the upload. Let me know if this works for you.

Mina

Hi Mina,

Yes that was one of the first things I noticed as a mistake I made, and so rectified that. What is the step for running an edit .biom file?

Thank you very much again.

Karl

Hi Karl,

When uploading your biom file you just need to change "Auto detect" in the "datatype" column to "picrust". Alternatively, if you have already uploaded your file you just need to click on: Edit attributes (looks like a pen) > Datatype  and then set "new datatype" to "picrust".

If I misunderstood your question what do you mean with "running an edit .biom file"?

best, Mina 

I guess there might be a problem with the format of my biom file.

I uploaded and allocated the attribute to picrust format.

The tried running normalise by copy number and it just fails to perform the analysis.

That is why I was asking about checking the file format using editing commands or function with the uploaded file.

The file came from Ion Torrent so I am not sure and may have to go back there or tto the tutorial.

Thank you again!

Sorry. I got you wrong. The biom format keeps changing and it's not backward compatible. The easiest way would be to convert your biom table into a tab separated text filt ande then chose "legacy QIIME format" in normalize by copy number 

That's what I thought. I just am not sure how to change it to txt file. Ie open in text editor and then save as tab delimited in the save options.

Would you recommend it like that, and then chose legacy format in the galaxy server?

if you have installed qiime, biom-format should be installed as well. 

type:

biom convert -i table.biom -o table.from_biom_w_taxonomy.txt --to-tsv --header-key taxonomy

more conversion can be found at

http://biom-format.org/documentation/biom_conversion.html

Thanks for the answer, now just where do I need to place the biom file for the conversion in Qiime? ie my docs for that command to run without having to edit anything in the command line or is it fine anywhere?

you need to cd into the directory containing your biom file and then replace the "table.biom" with your file name. If you want you can also send me your file and I'll convert it for you. I'll message you my email address

That's what I thought. Now I am just having trouble with my Virtual file vdmk file and creating a shared folder to mount, so I can dump the biom file there and cd to it.