finding structural variations without paired reads

[mike.t]

I have some 454 Titanium reads (not a paired lib, unfortunately) that I am aligning to the reference bacterial genome. I'd like to identify translocations, inversions and large indels, if possible using these data, but most methods that I've seen use paired reads for this. It seems to me that long reads (Sanger, titanium) could be useful for this too, if the read alignment is split, and each split aligns to two different locations in the reference genome. Does anyone know of a software that uses this approach or a paper that describes it?
Mike

[kmcarr]

Mike,

I have done this using the Roche gsMapper software. Even with single reads it will align them to a reference and then report variants (SNPs, indels and SVs). If you don't already have the software you can request it from Roche/454. Here is the contact page for software requests:

http://454.com/contact-us/software-request.asp

[HESmith]

Hi Mike,

A short methods paper describing split-read alignment for structural variation can be found here.

[mike.t]

Thanks to both of you. I've heard that Roche is very flexible about handing out their software to those who need it, so I just requested it from the link that kmcarr gave me. I'll also check out the paper that HESmith recommended