Regarding data analysis of Ion Torrent using QIIME

[Shree]

Hello All,
We are using Ion Torrent PGM as our sequencing instrument.

Can anyone please tell us if we can use QIIME for analysis of 16s rRNA
data generated using Ion Torrent PGM.

Also, how to use QIIME for single sample Library, because I was trying
to use QIIME for the same but I a unable to prepare Map.txt file for
one sample library.

[Tony Walters]

Hello Shree,

I haven't actually personally analyzed Ion Torrent data with QIIME, but I don't see any reason why one couldn't process the data with QIIME.

It appears that you can get data in SFF or fastq format.  If you got the data in SFF format, you would want to create fasta and qual files (http://qiime.org/scripts/process_sff.html?highlight=process_sff ) and then demultiplex with split_libraries (http://qiime.org/scripts/split_libraries.html?highlight=split_libraries).  If you have fastq format (should have one or two reads files, and a barcodes file) you would want to use split_libraries_fastq (http://qiime.org/scripts/split_libraries_fastq.html ).  The one unknown is how to interpret quality scores (we use 25 as a cutoff in split_libraries.py and a 'B' ascii code and below for low quality in split_libraries_fastq), as the calls may be different with the ion torrent software.  It's probably best to start with the defaults and see if a large number of sequences are discarded for low quality.

Once you're past the demultiplexing step, the remaining QIIME analyses should be the same as for 454/illumina data, as you have 16S sequences.

The mapping file should be able to handle a single SampleID and associated data (BarcodeSequence, LinkerPrimerSequence, Description).  Can you email me your mapping file so I can see how it's formatted?

Also, are you using QIIME 1.4.0 (latest release)?

Best regards,

Tony Walters

[Shrikant Bhute]

Thank you Tony for your quick reply.

Actually I do not have Map.txt file with me, in fact thats the reason I could not process my Ion Torrent Data.

So, I would like to know how to make a Map. tx.t file, especially for a single sample. Because we usually do not perform multiplex PCR on Ion Torrent, therefore we get the output in sff format but for individual samples. and i do not have any idea of making map.txt file for the same.

Thanks

--

Shrikant S Bhute

[Tony Walters]

Hello again,

Here are details about the format of the mapping file:  http://www.qiime.org/documentation/file_formats.html#metadata-mapping-files 

Because you have individual SFF files for each of your samples (that are not barcoded), it is a bit more work to get the fasta sequences assigned to the samples in QIIME-compatible format.

If there are details in the fasta labels that can distinguish each of the samples from each other, there may be a way to make this process easier.  If you type: head <path to one of your fasta files> you should get some sample lines; if you can post a couple of those, we can see if there is something in the labels that we can take advantage of to assign the sequences to your sampleIDs.

The other option is to create a mapping file for each of your sampleIDs, run split_libraries.py for each SFF (specifically the .fna and .qual file generated from the sff) and mapping file, with no barcode (-b 0 option).  I'm assuming the primer sequences are still present, so you can still remove those.

Best regards,

Tony Walters

[Jeff Werner]

The remaining question (for me) is: does denoiser work with ion torrent data? Despite the fact that signal is generated in a totally different way, you *should* get flowgrams as well as all the similar indel-type errors... I guess a I'm just intrigued by their system.

[Jose Carlos Clemente]

Even if it works out of the box, I would be dubious if the error
profiles are equivalent...for Titanium Jens adjusted the denoiser
profiles, so I guess at least that would be needed here as well, but
maybe he has some insights. Any takers for sequencing a mock community
with ion torrent?