search

hildebra / lotus2

Amplicon sequencing pipelines suitable for SSU (16S, 18S), LSU (23S, 28S) and ITS.
http://lotus2.earlham.ac.uk/
GNU General Public License v3.0
52 stars 17 forks source link

Same ASV assigned to different taxonomy in two samples #77

Open Amanda-Biocortex opened 2 months ago

Amanda-Biocortex commented 2 months ago

Hi,

Im looking to get consistent results per samples such that sample X in cohort A will have the same results as the same sample X in cohort B.

I have previously run dada2 (stand alone) on cohort A and cohort B and am now running the results through lotus2 lambda -taxOnly. Ive pulled the ASV taxonomy assignment for sample X in both cohorts and while the identified ASVs and counts are the same, the taxonomy results are different.

Do you know why the taxonomy of ASVs in sample X change dependent on which cohort the sample is in?

Apologies is this amounts to me not understanding lambda sufficiently!

hildebra commented 2 months ago

Hey, it seems a bit weird, since the tax should be independent of cohort. I suspect your dada2 ASVs are numbered differently between cohorts, or that they are not identical ? A quick check to make sure this is not caused by lotus2: Please run the same ASV file (eg from cohort A) two or three times through tax assign with lambda. This should always result in the same taxonomic assignments.