Dear Qiime2 developers,
I am facing a situation where Principal Coordinate Analysis (PCoA) is splitting my samples by sequencing runs. Therefore I would like to include sequencing run in the model to compare OTU frequencies in sample groups. Is it possible?
Hi @Polyana_Tizioto,
Could you please provide a bit more information about how your OTU tables were made for each run and how and when did you merged your data from the 2 runs? Did you you merge them prior to OTU picking or after?
I’ve seen this exact same scenario you describe happen in situations where OTU picking from each run has been processed separately and merged after.
With denoising methods such as DADA2 and Deblur, the sequence variants you get are unique and comparable across different runs, that means you can (and should in the case of DADA2) denoise each run separately then merge the feature tables after. In contrast, with OTU picking methods such as vsearch you want to merge your data before OTU picking as the arbitary OTUs can conflict between runs and create the artificial separation you observe.
2 Likes
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.