BowTie Accuracy

[Arupsss]

I am experimenting with BowTie v 1 for various length reads (75 to 200 bp). Now, my task is to find it's throughput and accuracy. But, I can't understand how to find it's accuracy. Now, I was suggested to use an fully accurate tool and compare it to BowTie in-order to find accuracy. Can anybody help me, is there any other method (because that is really time consuming work) ? Or if I must have to compare BowTie with another tool, can anybody help me which tool I can use for above purpose. Thanks in advance.

NB: BowTie uses only substitution for finding mismatch.

[dpryan]

One way would be to generate reads randomly from the genome and see if they map to the same place. Have a look at wgsim. Keep in mind that you're unlikely to get 100% accuracy with any aligner, since parts of the genome aren't uniquely mappable (depending on the genome in question of course).

[Arupsss]

One way would be to generate reads randomly from the genome and see if they map to the same place. Have a look at wgsim. Keep in mind that you're unlikely to get 100% accuracy with any aligner, since parts of the genome aren't uniquely mappable (depending on the genome in question of course).

I have checked that. Issue is BowTie matches in say 120 places on 100 read. So, I cant understand how to get accuracy for those 20 places. (I use 12 as an example)

[dpryan]

I have checked that. Issue is BowTie matches in say 120 places on 100 read. So, I cant understand how to get accuracy for those 20 places. (I use 12 as an example)

That's not an issue with any of the tools but, rather, how you seem to want to define accuracy. Particularly when dealing with sequencing errors or reads with SNPs, there will be parts of the genome that aren't uniquely mappable. What you're seeing is likely a result of that.