Burrito input from Picrust2

[PauObregon]

Dear all, I used picrust2 to make functional predictions on my 16S rRNA data. I am a little confused, as I have many outputs from the whole pipeline and don't know which ones are required for BURRITO to assemble the functional profiles. I am using these ones, but having errors in the BURRITO assembly (The following OTU_IDs are in the OTU table but are not present in the PICRUSt 16S normalization table):

• Feature table of the ASVs (in tsv), with the frequences in each sample
• KO_predicted count for each ASV (tsv)
• Metadata file (tsv) I've read in other post that the flag --stratified in picrust2 could be of help, but I didn't manage to assemble the BURRITO. Any help will be much appreciated!!

[engal]

What exactly is the format of the second file (containing KO_predicted)? If it's a prediction of genomic content (i.e. copy numbers of each KO encoded by each ASV), then you can use that with the "Custom genomic content table" option for the taxonomy-function linking method.

You can also use PICRUSt2's stratified output option as described here with BURRITO's "Function attribution table" option for the taxonomy-function linking method.

[PauObregon]

Dear engal, I mannaged to correctly assemble the burrito, Thanks a lot for your help!

[Theda-sys]

Hello,

So here is my problem: I am using the LotuS_SLV138 to process the data. Running PiCrust with this is no problem, and i am using the following commands to get the output: picrust2_pipeline.py -s clean-seqs.fa \ -i clean-rarefied-abun.tsv \ -o output \ -p 8 \ --stratified \ --per_sequence_contrib \ --verbose

Now I am very excited to use the BURRITO thingy, however I have no Idea what to use for the upload? And than which of the options to select: Is it possible to use KO_predicted.tsv for "Custom genomic content table" option for the taxonomy-function linking method? What do I have to use as Taxonomic Data, and is it even possible when the data was processed with LotuS??

I am really appricate help!! Thanks, me

[engal]

Hi,

You are correct, you can use the KO_predicted.tsv for the "Custom genomic content table" option, and you should be able to use the associated abundance table for taxonomy (i.e. the "clean-rarefied-abun.tsv" file you used for PiCRUSt input).

The primary component you may be missing is a taxonomy file which specifies the hierarchical taxonomic relationships between taxa (see example on upload page for formatting). Since you are not using Greengenes OTUs, you will need to supply such a taxonomy file with the "Custom taxonomy" option. If you have a very small number of taxa in your dataset, you could potentially use the "Custom taxon IDs with no hierarchy" option, but if you have more than ~20-50 taxa, this option is probably not ideal.

Best, engal