View Single Post
Old 07-24-2014, 03:56 AM   #2
lre1234
Senior Member
 
Location: US

Join Date: Aug 2011
Posts: 106
Default

While using a mapper to find matches is actually a nice idea, I am not sure that it would work. From my understanding (and i'm a little rusty on all of the details) of the CRISPR system is that it is not just the 23 nt sequence that is important, but also the few bp's that follow your sequence of interest also see (http://www.sciencemag.org/content/339/6121/823.full) . For example:

5'-YOUR23NTSEQUENCE-NGG-3'

Where the N is a wild card and can be any base followed by a GG. Essentially, you are looking for regions in the genome in which have your sequence (with or without some mismatches) followed by the NGG sequence. I would assume(?) that the R- program is taking this into account, but when mapping your not which may be why the mapping missed some sites. Perhaps, trying to artificially add the NGG to your sequence then re-map and see what you get might improve the system (not sure if this would work or if it is even possible).

Also, see here, in which they have a tool that can be used, but it might be a little cumbersome with 3 million sequences:
http://www.genome-engineering.org/crispr/?page_id=159
lre1234 is offline   Reply With Quote