Facing issue for importing representative sequence file into picrust2

[Rishikesh Dash]

Hello Picrust community,

I have representative sequence files forn each sample in fasta format. But As per picrust command it require one single representative sequence file which include sequence id as well as sample information. Please help me how to import my representative sequences into picrust??

[Robyn Wright]

Hello,

Have you first processed your raw files? e.g. using the QIIME2 pipeline? You will need to have a file containing representative sequences as well as one containing counts of these sequences in your samples. Is this what you have for each sample? 

Best wishes,

Robyn

[Allan Amorim Santos]

Hello everyone.

have you got any idea how should I proceed using outputted files from Mothur?

[Robyn Wright]

Hi there,

I haven't used Mothur recently so I'm not 100% sure what your final output looks like, but all you need for running PICRUSt2 is a fasta file containing representative sequences (it doesn't matter whether they are ASVs or OTUs) and a tab-delimited or biom file containing the abundances of these sequences across your samples, as described here. It doesn't matter which pipeline (QIIME/Mothur) has been used for generating these files. 

[Allan Amorim Santos]

Hi Robyn.

I got it. However, I found some new doubts.

I run v4 samples following the pipeline resulting in the last fasta file after trimming, aligning against silva 138_1, filtering, removing chimeras, etc.
However, in this last fasta there were so many gaps (-) which I believe carried picrust running to an error type "Stopping - all 100 input sequences aligned poorly to reference sequences (--min_align option specified a minimum proportion of 0.8 aligning to reference sequences)."

I've removed all these gaps and then run again;
It has worked after running the standard command "picrust2_pipeline.py -s study_seqs.fna -i study_seqs.biom -o picrust2_out_pipeline -p 1"

However, I'm not sure if I'm right in removing these gaps from the last fasta.
Additionally, I've selected the top 50 most representative ASVs followed by their respective sequence in fasta.

Could anyone tell me it I'm right or not after carrying this?
Thank you so much!!

[Robyn Wright]

Hi there,

Apologies for the slow reply. You're right to remove the gaps - the fasta that you input into PICRUSt should be an unaligned fasta, and the gaps come from your alignment to the SILVA 138 database. 

Robyn