Thanks for the debug log @Yalu_Zhou! This is the important part of the error message:
Error in isBimeraDenovoTable(unqs[[i]], …, verbose = verbose) :
Input must be a valid sequence table.
This error occurs if none of the forward and reverse reads can be joined during the denoising process. You’ll need to check that your forward and reverse reads overlap enough to be joined after applying the read trimming specified with --p-trim-left-f
, --p-trim-left-r
, --p-trunc-len-f
, and --p-trunc-len-r
. After trimming, the reads must overlap by at least 20 nucleotides plus the biological length variation of overlap (see @benjjneb’s posts here, here, and here for details).
Additionally, you’ll need to ensure that all sequencing artifacts (e.g. primers, barcodes, adapters) have been removed from your reads prior to denoising. If the primers are contained at the start of your forward reads, you can alternatively use --p-trim-left-f
to have DADA2 trim off the forward read primers during denoising.
The quality score plots generated by qiime demux summarize
can be helpful with choosing appropriate trimming parameters based on your sequence data. Hope this helps!