Consistency of feature Ids after merging rep-seqs.qza files

User Support
1

Hey there,
So I have a concern. I'm trying to merge rep-seqs.qza files, but I have seen that qiime2 looks out for features with the same sequence, merges them, and gives them new feature IDs. Is this observation true?
If this is true, then it raises a problem, since the corresponding table.qza files have different feature IDs and will not recognise the new feature IDs in the merged-rep-seqs.qza file.
Unless, qiime2 has a system where particular sequences are always given particular feature IDs rather than them (the feature IDs) being generated randomly. I do not know if this is the case.
Please, I need urgent responses to know if I can demultiplex my samples separately and and can merge them later, or if I must always demultiplex them together to avoid the above uncertainty.

Thank you

2

Hello!
You can denoise your samples in batches and then merge representative sequences and feature tables! Make sure to run cutadapt and Dada2 with absolutely identical parameters to assure compatibility between batches.

Same sequences always have the same IDs since ids are md5 hashes of sequences themselves and not random.

Same sequences in different batches will be merged without changing their ids.

Best,