Dada2 back to qiime 2

User Support
11

Hey all!

I did a bit of research on this, and it turns out there’s an easier way!

I’m going to use the DADA2 1.4 Tutorial as a reference. In order to accomplish things like beta diversity analysis we’re going to need the sequence table (we call it a feature-table), and the sequences. In the tutorial, those can both be made from seqtab.nochim which I will reference in a few places.

I’m also going to imagine we have a directory dada2-analysis/ which I’ll use in place of an actual filepath.

In your R session, you’ll want to run the following (replacing the particular filepaths):

write.table(t(seqtab.nochim), "dada2-analysis/seqtab-nochim.txt", sep="\t", row.names=TRUE, col.names=NA, quote=FALSE)

uniquesToFasta(seqtab.nochim, fout='dada2-analysis/rep-seqs.fna', ids=colnames(seqtab.nochim))

We can import that fasta file easily with:

qiime tools import \
  --input-path dada2-analysis/rep-seqs.fna \
  --type 'FeatureData[Sequence]' \
  --output-path rep-seqs.qza

For the feature-table, there are two steps we have to do first:

Add a special header for BIOM:

echo -n "#OTU Table" | cat - dada2-analysis/seqtab-nochim.txt > dada2-analysis/biom-table.txt

Convert to BIOM v2.1:

biom convert -i dada2-analysis/biom-table.txt -o dada2-analysis/table.biom --table-type="OTU table" --to-hdf5

Now we can import that as well:

qiime tools import \
  --input-path dada2-analysis/table.biom \
  --type 'FeatureTable[Frequency]' \
  --source-format BIOMV210Format \
  --output-path table.qza

This should leave you with a rep-seqs.qza and table.qza that you can use (following along with the moving pictures tutorial).

11 Likes