I had the same sample sequenced with same barcode in two runs. So I got two sff file for the same sample. when I want to denoise the sequences, Should I firstly merge the sff files together and then denoise once, or denoise them separately and merge them later?
Appreciate for any suggestion!
I would run the SOP on each run separately through chimera checking, then merge
Actually… you would want to pool them after trim.seqs to make sure they are aligned the same way. We actually have a command that does this for you: sff.multiple (http://mothur.org/wiki/Sff.multiple).
Thanks！ That helps a lot
The problem is that some sff file contain few sequences (~5000 sequences). For example, Sample 1 has two sff files. One contains 35000 sequences, the other contains 5000 sequences. As I know shhh.flows will denoise each flow file separately. Will denoising with shhh.flows command be effective with such small data sets?
ah right forgot to say that I realigned them after merging. Chimera checking can be computationally expensive, I didn’t want to repeat that step for all samples each time I got a new batch of sequences