Joining Paired-End Illumina Raw Reads

Question

[Yongjie Zhang]

I recently have amplicons sequenced (Illumina PE250) to investigate microbial community. I want to know what quality detection, and what kind of trims of raw reads by what programs should be performed before joining paired-end raw reads? I think low-quality parts, potential barcode or primer sequences existed in raw reads should be excluded in advance. Is this right?

[Yongjie Zhang]

I finally merged reads pairs using usearch -fastq_mergepairs with -fastq_allowmergestagger enabled, but before merging, I cut primer/adapters using CutAdapt and trimed low-quality regions from ends of reads using Trimmomatric, and after merging, I cut potential primer/adapter again by CutAdapt and filtered reads with >0.5 expected errors by usearch -fastq_filter. Then I can do downstream analyses.

Answer 1

[Janake]

Here is a good example to start: http://www.mothur.org/wiki/MiSeq_SOP

[Yongjie Zhang]

Very useful link. Thanks. I found the join_paired_ends.py in Qiime is not very ideal to my data. Mothur might be better.

Answer 2

[Danielk]

You could use SeqPrep for this. see https://github.com/jstjohn/SeqPrep or my fork https://github.com/dakl/SeqPrep which adds two parameters for more stringent handling of mismatches (set to N, essentially).

Answer 3

[Joseph Hughes]

You should be able to find everything you need for removing adapters, splitting the sample based on barcodes and joining the pairs with a combination of the tools in the fastx_toolkit and ea-utils.