Hi! I am using DADA2 and phyloseq in R and I get slightly different results if I run the commands more than once with the same sample. I am working with 16S rRNA gene amplicons (Illumina, single end reads). Is there some randomness involved in this pipeline?
https://ryjohnson09.netlify.app/post/microbiome-analysis-with-dada2-and-phyloseq/
Any help would be appreciated! Thank you!
Hi Camilla,
It happens if you change the quality control trimming criteria (number of bases to trim). Are you doing that? Otherwise, I have not seen such a thing.
I am using DADA2 very often.
Best,
I guess you got the answer from Ryan.
However, what differences were you getting? And on what level?
Yes, I got the answer from Ryan: "the error rates/inferring sequences does involve a bit a predictive modeling (if I remember correctly) which may vary slightly from one prediction to the next."
Then I confirmed this is correct by reading the publication: http://www.nature.com/articles/nmeth.3869
The difference were very subtle (a few reads) on the level of genus.
Thank you!
- Badges
- Science method