question about creating featured table

Technical Support
1

I have fastq.qz files that is prepared as described below,
I was able to import it to qiime 2 and denoised it again using dada2 to create featured table, my question is how can I create featured table without going through dada2 since my files are already denoised if I understood it right.

FASTQ files
16S amplicons from each sample are individually barcoded and sequenced in multiplex in the NextSeq 500 platform in a 150bp paired-end modality. Raw data from the sequencer is first demultiplexed, and the forward and reverse reads obtained in each of the 4 lanes per sample are filtered using the following criteria:

  • both forward and reverse reads in a pair must have an average Q-score > 30
  • primers, and any leading random nucleotides (used to increase diversity of the library being sequenced) are trimmed, and forward reads are capped at 125bp and reverse reads are capped at 124bp
  • forward and reverse reads of each pair are appended, and those sequences that contain more than 8 consecutive same nucleotide repeats are discarded
  • the remaining sequences are clustered using a distance of 1 nucleotide using the Swarm algorithm1, and the most abundant sequence per cluster is considered the representative of the cluster and it is assigned a count corresponding to the sum of sequences that compose the cluster
  • a chimera removal using these centroid representative sequences is performed using the VSEARCH uchime_denovo algorithm2. Singletons that remain after chimera removal are also discarded.
  • Finally, both forward and reverse reads that match with at least 77% sequence identity to the same sequence in version 123 of the SILVA database3 are assumed to be 16S sequences

Below is how the Fastq file looked like:

@NB501656:349:HW37GAFXX:1:11101:18457:5769 1:N:0:ACTGTGGA+GTGAGGAG
TACGAAGGGGGCTAGCGTTGTTCGGAATTACTGGGCGTAAAGCGCACGTAGGCGGATCGATCAGTCAGGGGTGAAATCCCAGGGCTCAACCCTGGAACTGCCTTTGATACTGTCGATCTGGAGTA
+
AEEEEEEEEEEEEEEEEEEEEEEEEEEEEAEEEEEEEEEEEEEEEAEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE/EEEEEA/EEEEEEEEEEEEEEEEEAEE</<EEEEEE
@NB501656:349:HW37GAFXX:1:11109:8309:12671 1:N:0:ACTGTGGA+GTGAGGAG
TACGTAGGGGGCTAGCGTTATCCGGAATTACTGGGCGTAAAGGGTGCGTAGGTGGTTTCTTAAGTCAGAGGTGAAAGGCTACGGCTCAACCGTAGTAAGCCTTTGAAACTGGGAAACTTGAGTGC
+
AEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE/EEEEEEAAEEEEEEEEEE/EEEEEAEEE<EAAEEEEAEEEAEEEEEEEEEEAEEEE
@NB501656:349:HW37GAFXX:1:11112:24082:5477 1:N:0:ACTGTGGA+GTGAGGAG
TACGTAGGTGGCAAGCGTTGTCCGGAATTATTGGGCGTAAAGCGCGCGCAGGTGGTTTCTTAAGTCTGATGTGAAAGCCCACGGCTCAACCGTGGAGGGTCATTGGAAACTGGGAGACTTGAGTG
+
EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE<E6EEEEEEEAEEEEEEEEEAEEEEEEEEA/EEEEEEEEEAAAEEEEA
@NB501656:349:HW37GAFXX:1:11112:8576:7759 1:N:0:ACTGTGGA+GTGAGGAG
TACGTAGGGTGCAAGCGTTAATCGGAATTACTGGGCGTAAAGCGTGCGCAGGCGGTTATGCAAGACAGAGGTGAAATCCCCGGGCTCAACCTGGGAACTGCCTTTGTGACTGCATGGCTAGAGTA
+

2

Hi @Dawud922,
Sounds like filtering and clustering of your raw reads has been done outside of Qiime2 and now you just want to import your processed feature-table into Qiime2 for further analysis. Usually by the end of these processes you will have a feature/otu table, typically in .biom format or perhaps a text file. You’ll want to import that instead of your raw fastq files. The raw fastq files are imported when you want to do denoising/clustering etc in Qiime2. Since you’ve already done those steps elsewhere, look for a biom (or text) file of your feature table. Check out the importing tutorial for an example of how to import this type of file.

3

I don’t have the .biom format file. After importing my fastq.qz file I denosied it again using DADA2 and seems works fine, I can run rest of the codes to do analysis. What do you think?
Thanks

4

Hi @Dawud922,
Yes that is perfectly acceptable (and probably even better) than the protocol you mentioned. Just so you know though then that you are starting your analysis from scratch with raw fastq files, so the list of steps you mentioned in your post are no longer being used. I mention this so when you are reporting your methodology you would be basically ignoring those.

5

Great, It’s always nice to have others advices.
Thank you so much.

1 Like