From merged reads from pandaseq, can I start demultiplexing with QIIME2

demux
(Ele) #1

Hi,

I am trying to obtain 16S diversity form my samples. From my raw data, I have trimmed and merged my forward and reverse reads with cut adapt and pandaseq. My outputs are now .fasq files. Could I now start demultiplexing using “qiime demux amp-single” as input has to be in .qza I am not sure how to start from there. I am working with Paired ends, how can I continue from here?

Thanks,

(Justine) #2

Hi @ecg

I found the easiest way to do this was to build a manifest, and then import the sequences as as a “SampleData[JoinedSequencesWithQuality]” type. My import command for vsearch joined sequences looks like this:

qiime tools import \
 --input-path ./manifest \
 --output-path ./ready_seqs.qza \
 --type 'SampleData[JoinedSequencesWithQuality]' \
 --input-format SingleEndFastqManifestPhred33

where I defined my $manifest_path and $art_path as bash variables.

Then, I suggest following the Alternate methods of read joining tutorial from the Viewing a summary of the joined reads step. Please note that you can’t use Dada2 with pre-joined reads, since Dada2 uses read joining as part of their quality filtering, where as Deblur will accept it.

Best,
Justine

(Ele) #3

Hi Justine, thanks for the quick answer.

I have created the manifest and run
qiime tools import
–input-path manifest.csv
–output-path ./ready_seqs.qza
–type ‘SampleData[JoinedSequencesWithQuality]’
–source-format PairedEndFastqManifestPhred33

using Paired end instead of single reads as I have both forward and reverse per sample, but Ihave the following error:

An unexpected error has occurred:

No transformation from <class ‘q2_types.per_sample_sequences._format.PairedEndFastqManifestPhred33’> to <class ‘q2_types.per_sample_sequences._format.SingleLanePerSampleSingleEndFastqDirFmt’>

Do you have any idea what does this mean?
thanks
E

(Matthew Ryan Dillon) #4
(Matthew Ryan Dillon) #5

Hi @ecg!

You mentioned above that these reads are pre-joined, which means the --source-format should be SingleEndFastqManifestPhred33, as @jwdebelius mentioned (not PairedEndFastqManifestPhred33).

This seems to be at odds with this statement though:

Are you talking about two different datasets here? If they are pre-joined (what you called “merged”), then you wouldn’t have a forward and reverse read per sample. Can you please clarify? Thanks!

(Matthew Ryan Dillon) #6
(Ele) #7

Hi @thermokarst

Apologies for the confusion, I started working with the joined (already merged samples), but then I have read from other forum posts that in order to use dada, and quality filtering it is advisable to work with the initial forward and reverse reads.

Doing so, I have decided to start with the initial forward and reverse form the different samples, using:

qiime tools import
–input-path manifest.csv
–output-path ./ready_seqs.qza
–type ‘SampleData[PairedEndSequencesWithQuality]’
–source-format PairedEndFastqManifestPhred33

And got this error:

An unexpected error has occurred:

**Forward and reverse reads must be provided exactly one time each for each sample. The following samples had forward but not reverse read fastq files: ECsample1_R1.fastq.gz, ECsample2_R1.fastq.gz + ALL THE SAMPLES

Is there a problem on the manifest? I have checked names and paths and it seems to be ok.
Should I check something else?
Thanks

E

(Ele) #8

Hi,

I have re-started one more time, and got all working.
I am not sure yet what I had wrong but I have solved it.

Thanks again.