250bp paired end Miseq reads alignment

[empyrean]

Hello all

Recently we upgraded our miseq and got 2*250 run data. Does anyone have experience in aligning this sort of data? I have been using bwa till now for previous runs when it produced 2*150. Unable use bwa on this data using Bwasw as it didnot support paired end data. I am looking in to bowtie2 now to use as aligner. My references are amplicons and the size is approximately 2kb. In bowtie2, they are suggesting to use bowtie2 for longer genomes. Any thoughts on this?

[krobison]

One approach would be to use a tool such as FLASH to merge many of your read pairs into single reads, then align those long reads with bwasw.

Of course, for what doesn't merge you still have the old problem. But for very long reads, and especially matching to a small target genome (amplicons), I'm wondering if just aligning them independently with bwasw will work just fine. Paired-end alignment is convenient and may give better specificity in some cases, but a large fraction of the time, particularly with a small genome, it isn't actually gaining anything. Of course, you will lose the convenient auto-generation of insert size information, and some sort of post-processing would be needed if you require that.

[JackieBadger]

you can use Bowtie/Bowtie2...BLAT...BFAST

[swNGS]

The following paper suggests that bowtie2 offers a marginal advantage over bwasw when it comes to aligning 2x250 versus 2x150. They did use simulated data however, but worth experimenting with.

http://www.nature.com/nmeth/journal/v9/n4/abs/nmeth.1923.html