I imported paired-end reads after a quality filter with trimmomatic in qiime2 (using casavana). I’m wondering if it’s possible to generate a table.qza file without passing the data into another quality filter process.
I had made this process because each time I filtered my samples in dada2 I got the error “return code-1”. I think I got that since I’m working with V1-V3 reads.
If your data is pre-quality filtered, it is not possible to use dada2. Dada2 relies on the quality scores to build a profiling model and so your sequences cannot be quality filtered already. You can try passing into deblur. The quality filtering you used has not been bench marked as part of that protocol, but it is a more appropriate approach.
It’s no possible to generate the file table.qza without using any additional quality filtering step? (I don’t want to use dada2 nor deblur, just use the sequences as they are because they are already filtered).
If trimmomatic is giving you a feature table, then you can import the data as a feature table. Otherwise, you need to apply an algorithm to convert to a feature table - either you denoise to a amplicon sequence variant (ASV) which is an endpoint in and of itself, or if you’re feeling like it, you could either denoise or go directly into OTU-based clustering. I discourage the second because you’re giving up a lot of information in favor of historical precdent. In your case, i would run your quality filtered data through deblur and then stop there as a feature table. But, to turn your demultiplexed data into a feature table, you either need to denoise or cluster.
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.