Need some advice as to why Archaea PICRUSt functions present in QIIME2 data with no Archaea identified

[BioHazard Dragon]

Hi, I'm trying to work out what exactly is happening with my workflow. Hopefully, someone can give me some advice to ID what is happening. 

I have samples that were run through QIIME2 and produced a Feature Table and Taxonomy (using SILVA 138.1) that did not identify any Archaea (I was expecting a few but none were identified). 

I used the files produced using QIIME2 to create a Phyloseq object and cleaned up the data (orphan ASVs, singletons and doubletons). 

I extracted the tables from the filtered Phyloseq object and used them as input for PICRUSt. When I checked the inferred MetaCyc pathways (path_abun_unstrat_descrip.tsv) against the MetaCyc database - I had a number of paths that were identified as specific to Archaea. 

Does anyone know why this is happening and if there is any way to identify which ASVs were mapped to these Archeal-specific pathways? 

Cheers

Rachele 

[BioHazard Dragon]

I forgot to add that the fasta file used as input, was filtered to match the filtered Phyloseq so that the ASVs were the same.

[Gavin Douglas]

Hi there,

The MetaCyc pathways are based on what gene families are predicted to be present. These gene families are often very broad and are found across both bacteria and archaea, even though they may be involved in very different pathways in each lineage. So it’s most likely that those are false positives (but note you would likely receive the same false positive with shotgun metagenomics data too, if the gene families involved in the pathway are actually present, even if no archaea are present).

Cheers,

Gavin

-- 
You received this message because you are subscribed to the Google Groups "picrust-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to picrust-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/picrust-users/9270cf45-945c-4e7f-80e2-ae04f7f8ca8bn%40googlegroups.com.