I am planning on running qiime's pick_open_reference_otus.py on a mixed set of 300-400 16s samples, some of which are Illumina MiSeq and some are 454. At this point I have only fasta & fastq data, I don't have the original sff files for the 454. They are all already demuxed and I have 1 file per sample.
I want to use this data to build genera level OTUs on a per-sample basis and was wondering if denoising is really needed in this case. If I were working at the species level I can see how it would be needed, but at the genera level I'm not so sure it would help. Also I'm not sure whether its even possible if I don't have sff files. And I'm not exactly sure how to denoise MiSeq data if I do need to do this. Does anyone have any suggestions or advice?
Thanks,
John Martin
I want to use this data to build genera level OTUs on a per-sample basis and was wondering if denoising is really needed in this case. If I were working at the species level I can see how it would be needed, but at the genera level I'm not so sure it would help. Also I'm not sure whether its even possible if I don't have sff files. And I'm not exactly sure how to denoise MiSeq data if I do need to do this. Does anyone have any suggestions or advice?
Thanks,
John Martin