triming process and saving trimmed reads

[ramiroricardo]

Dear all,

We are trying to se Lotus2, with dada2, to analyse some Illumina 16S data. We have two questions:

• When using dada2 directly, one would typically have to define the size of the reads for trimming, for example triming 250bp reads to 230bp, which would be applied to all fwd/rev reads. With Lotus2, we are setting TrimWindowWidth=20 and TrimWindowThreshold=25. My question is whether this is if the same length is then applied to all reads, as with dada2, or if we have different lengths depending on the specific quality of that read?
• If we want to save the quality trimmed reads, would we do it through the "-saveDemultiplex" option? (asking as the documentation mentions the quality filtering but not trimming)

Thanks for any help,

Ramiro

[hildebra]

Hey Ramiro, the length of sequences in dada2 is indirectly controlled in LotuS2. LotuS2 has its own sequence filtering algorithms implemented in "sdm". This is done for a) speed and b) sdm offers many more advance quality filtering criteria. I see that you already edited the sdm config files. The parameter "minSeqLength 200" and "TruncateSequenceLength 200" together control that sequences will be cut after 200 bp, you can change this parameter, in fact it is recommended to adapt this parameter to every experiment (there are several sdm related log files that can give you statistics how many reads are at what qual at different lengths etc, please have a look). If you want to export the qual filtered reads, two ways 1) "-saveDemultiplex 2" 2) (to make sure you really replicate what's happening): abort while dada2 is running within LotuS2, it will give you the command to our R script interface with dada2, it might be easiest to load this script and run it line by line to replicate LotuS2 output

Last, there is of course the option to use "lotus2 -i ... -CL dada2" that should do all that automatically.

hth, Falk