How to demux the two same barcode database

heavennew · January 23, 2018, 9:50am

Hi,
I have two database of sequences. One had sequenced last year, and the other sequenced this year. I want to combine these two database to analyze.But both of them had same barcode. Qiime2 unable to demultiplexed the same barcoded database. So how can I combine these two database to complete my analysis with qiime2?

Micro_Biologist · January 23, 2018, 11:02am

You will probably want to run these through analysis separately to begin with.

Are you using dada2? If so importing each run individually, running each dada2 analysis and then merging tables and rep-seqs will give you what you want.

heavennew · January 23, 2018, 11:26am

Hi, @Micro_Biologist
I used deblur, but I think the way that you suggested may contain the many replicate sequences (feature sequence).
How do you deal with it?

Micro_Biologist · January 23, 2018, 12:59pm

Okay, truth be told I am unfamiliar with the deblur pipeline.

Having said that it look like to me you will have to import them into demux.qza files and run deblur separately. and then merge them using something like:

qiime feature-table merge \
--i-tables Table1.qza \
--i-tables Table2.qza \
--o-merged-table MergedTable.qza

qiime feature-table merge-seqs \
--i-data RepSeqs1.qza \
--i-data RepSeqs2.qza \
--o-merged-data MergedRepSeqs.qza

If my understanding is correct deblur’s features (similarly to dada2) are derived from the nucleotide sequences and so Features should be the same across runs.

heavennew · January 23, 2018, 1:31pm

Hi, @Micro_Biologist
Yes, the code are right. But I just guess that this way may combine the same sequences (replicated feature sequences) .
I think it is a big problem.

Micro_Biologist · January 23, 2018, 1:42pm

My colleague has just informed me that Feature IDs are the same across runs as they are derived from the nucleotide sequences themselves.

So long as your settings for deblur are the same the 2 tables and rep seqs will merge without an issue, as the feature IDs will be consistent between runs.

If its at all possible to try that it should work!

heavennew · January 23, 2018, 2:43pm

Hi, @Micro_Biologist
It is so good that the Feature IDs are unique for each sequences.
Thank you for your reply.

wasade · January 23, 2018, 7:58pm

I agree with @Micro_Biologist

Best,
Daniel

system · February 24, 2018, 2:55am

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.