Can LotuS pipeline with sdm generate stitched reads in output directory for closed-reference OTU picking in QIIME???

[qinglong wu]

Hi Falk,

I got demultiplexed paired-end reads in hand, but I also need stitched reads for closed-reference OTU picking in QIIME. Since sdm pipeline could demultiplex, filter reads and remove primers/barcodes perfectly, I was thinking wehther we can generete stitched/merged reads for each sample during every LotuS run.

I read the information as below in the web documentation, but in the output folder, I can not find the merged reads.

Link: http://psbweb05.psb.ugent.be/lotus/UsageSamples.html   

"LotuS will take care of further processing of these files from here including merging of the paired reads and can be started using the following command:

Please note, that I recommend to let LotuS handle the read merging, as the reverse read has on average ~5 lower quality and needs to be treated with care. However, LotuS was adapted to this and tries to get the maximum phylogenetic information per OTU, merging the second read to the first after high quality OTUs were built from the first, high quality read."

I think I did not obtain from current LotuS run. But how can I have access to the merged reads as output for miseq paired-end reads?

Look forward to receiving your comments.

Thanks.

Qinglong

[Falk Hildebrand]

Hey Quinglong,

this is a bit more complicated, but in short: lotus is only so fast (a lot faster than mothur or qiime), because it only does what is absolutely necessary. This also includes on stitching reads which need to be stitched and this is the OTU representative seqeunce that needs to be long in order to assign the taxonomy with the best resolution.

In your case, I would recommend to run flash on the demultiplexed output files yourself, flash is in the lotus dir. Note that the lotus output assigned to species level will give you a similar resolution to closed ref OTU picking, but you need to remove unassigned taxa in this case from the species level table. If you want to get your OTUs assigned to a specific taxonomic tree (like the Qiime proveded one), using QIIME closed ref OTU picking is a viable choice, though be careful not to overestimate diversity with this.

best,

Falk

[qinglong wu]

Thanks, Falk!

Yes, I will think the closed-reference picking carefully in the future.

Much appreciated for your quick answer.

Nice weekend.

Qinglong