Hi. I would suggest that you use dada2 for data analysis. You have 2 options: first you can inspect both reads and use only one which has a better quality (SE, most likely fwd) or you can use the justConcatenate flag in the merge pairs step to join them sep by Ns. 

http://benjjneb.github.io/dada2/.html

QIIME can be used for this purpose. It can take raw reads (without stitching) as input and map them against various databases to identify the OTUs using LCA algorithm and assign relevant taxonomic relationship among these OTUs.

So stitching (on the basis of overlap in R1 and R2, as not available in your case) is not require any more. 

Hope it helps ..

Thanks!

do you merge them? could you provide me an example of the command.py?

Usually PE data for amplicon sequencing is generated with 2x 300 bp chemistry. Length of targeted amplicon (variable regions V3-V4 usually) vary between 500bp to 550bp, So in this case PE reads can be stitched to represent a single amplicon sequence.

I use PANDASeq or FLASH to perform this task, you can also use join_paired_ends.py given in the package of qiime.

Hi. I would suggest that you use dada2 for data analysis. You have 2 options: first you can inspect both reads and use only one which has a better quality (SE, most likely fwd) or you can use the justConcatenate flag in the merge pairs step to join them sep by Ns. 

http://benjjneb.github.io/dada2/.html

Hi Luigimaria. Take a look to the pipeline we developed at Wageningen University, no need of overlapping reads.

http://f1000research.com/articles/5-1791

@Javier Ramiro-Garcia,

It looks like your website is down, can you please send me your username/dataset/code and along with the article. Thanks