question about order of analysis events in Lotus + answer

[jobl...@gmail.com]

Maybe interesting for some of you:

I am trying to understand how the Lotus pipeline is performing tis different steps compared to other analysis tools. Is there maybe a document or manual where I could find in which order the analysis steps in Lotus are performed? 

eg. starting from already demultiplexed non-merged paired end reads:

 1) quality filtering / trimming (is quality filtering performed on trimmed reads or after?? (TruncateSequenceLength option, TrimWindow option and trimming due to reaching the maxAccumulatedError)

2) singleton removal : are these re-entered for upon counting of OTUs (like the mid-quality sequences)

3) dereplication

4) Uchime

5) OTU clustering (I was confused by the manual: are all high quality reads included or only the high quality reads from the forward reads)

6) Seed extension: from what I understand from the manuscript, merging of fwd and rev reads only happens at this step and only for the selected OTU representative sequences. Is that correct?

7) at this point mid-quality (and singleton) sequences are clustered in the existing OTUs and abundance is estimated

Does this seems correct to you?

[jobl...@gmail.com]

Roughly steps happen in this order:

1-3 is done in sdm

5+4. only high quality, derep passed reads are used, fwd read only -> this is due to increased OTU clustering quality you get from using fwd read only

6. yes, only the highest quality, longest representative read pair is selected among all reads in this OUT. This one is then merged.

4. (ref based)

7. post processing: mapping singletons and mid qual reads on the “good OTUs” in order to max counts recovered, counting OTUs per sample, removing contaminants, building phylo tree etc

 

Hth,

Falk