FastQ quality trimming/keeping the reads paired

[mstagliamonte]

Hi everyone,

my name is Max and I'm a newbie in de novo assembling genomes. I have often read this forum (+ a few tons of literature and I am trying to step up a bit now.

I am trying to use Velvet/Metavelvet on some fastq paired ends data, and I would like to do some quality trimming on my data before feeding it to the assembler. The problem is that Velvet requires all reads to be in order and paired if the input is a fastq/fasta file.
If I filter the reads by quality, I am afraid not all the reads would be paired anymore.

Would a conversion of the unaligned reads into a sam file keep the information regarding pairing? Velvet does work with sam, and in this case it does not requires all the reads to be paired. I have read a bit about sam format, but I am not sure what I'm saying makes sense... does it?

Thanks for all the time you spend answering and helping on this forum!
Max

[westerman]

If I filter the reads by quality, I am afraid not all the reads would be paired anymore.

There are many posts on this forum about how to keep pairing when doing quality filtering. Most of the posts talk about Trimmomatic which is a good keyword to search on.

Would a conversion of the unaligned reads into a sam file keep the information regarding pairing?

Yes the pairing information can be kept. The sam format has flags as to if 'the read is paired in sequencing' and 'the read is the first read in a pair'.

[mstagliamonte]

Thanks!

Sorry for posting something that had been already asked. I'll try to do a more extensive search next time


Have a nice day
Max