Whole genome alignment (or the opposite of it?)

[jomaco]

Hi

We have two preliminary genome assemblies (which are fragmented libraries of contigs/scaffolds comprising sequencing data from two different types of the same species of plant). We believe the locus we are interested in contains large lengths of sequences only present in one of the types and completely absent in the other.

I have looked at various whole genome alignment programs but these seem more suited to looking at the similarities between two assemblies and not the differences.

I could use exonerate/BLAT to align all of the sequences of one plant to the other. The contig IDs which aren’t hit could then be used to pull out type-specific sequences from the corresponding assembly.

I think the problem with this approach however would be that if there is a repetitive region in a contig this will be hit; even if the part of the contig is one of these type-specific sequences. I could write a program to look at the positions that are hit rather than the contig IDs but I just thought I’d check to see if I am missing something in the exonerate/BLAT approach or if there is a tool out there to do this.

Thanks

Jom

[turnersd]

Something like this what you're looking for?

http://mummer.sourceforge.net/