Mate pairs processing

[Antony03]

Hi,
I will receive soon mate pairs reads from illumina MiSeq. I will have the circularization adapter and need to trimm it (with cutadapt) from reads before assemble them de novo. Do I need to reverse complement them before the trimming step?

After, I will need to remove orphan reads, what do you think about the script filteresCommonPEreads.sh in this page:http://seqanswers.com/forums/showthr...oto=nextoldest?
Thanks!

[sphil]

You don't have to reverse complement. Assemblers are aware of PE sequencing, if you forward the right flag of how the sequencing is done. Can't comment on the scirpt, first guess: give it a try...

cheers

[winsettz]

I've /heard/ that velvet required reverse-complementing; but generally if an assembler has explicit flags to set mate-pair libraries, it should generally be able to handle them properly without reverse complementation (like SPAdes, for example).

When you say "orphan reads", you mean reads without their mate? If your workflow includes quality-based trimming of reads (eg Trimmomatic), Trimmomatic will process paired reads and sort out R1 & R2 into 1P/2P (reads with mates) and 1U (1 unpaired) and 2U (2 unpaired).

[BugSeq]

I would recommend using filtering programs that keep track of the mate pair relationship and separates them into another file when the mate doesn't meet your filtering requirements. I have good luck with Trimmomatic (http://www.usadellab.org/cms/?page=trimmomatic).

In regards to the manually reverse complimenting your data, that will probably not need to be done with most assembly programs. If they have an explicit flag or similar than the program will handle the reads properly. It depends on the assembly program. So pick your poison and read the manual.

Assuming no other type of data and depending on the size of the genome being assembled, you may want to check the diversity or your reads before assembly. You may not have the coverage you think.

[winsettz]

Also, the "reverse complement" stuff comes from the illumina nextera application note. The note is "Current as of 17 December 2012" (Pub. No 770-2012-053).

Prior to performing de novo assembly,the data were pre-processed as follows. In short, the adapters were trimmed and the RF reads were reverse-complemented. Additionally, the data were randomly downsampled to 500 MB to limit mean coverage to 50-100x, thereby reducing the computational requirements of the assembler.

[Antony03]

I want to use the assembler Ray and I need to reverse complement the reads.

Trimmomatic looks nice. But I'm not really sure how make 'The Adapter Fasta' for cut the circularization adapter.