polymorphisms among mouse strains concern for mapping the reads from ChIP-seq?

[ramkumar]

Given that the reference genome to map the sequence reads is usually of C57BL/6, do you care about the genetic background of the animal from which you get cells for your ChIP?

[swbarnes2]

I think the answer has to be "yes, it will make a difference". There might be some regions where your sample will differ too much from B6, and reads will fail to map, so you will miss that peak, which is exactly what you do not want.

You might try de novo asembly on unampped reads, and then figure out where those contigs align.

[ramkumar]

Thank you very much swbarnes2

[mudshark]

it will make a difference or sure. but i am in doubt whether it really matters and whether it is really a strong effect.

a) you will need quite some polymorphism to fail any read mapping in a peak region - given a peak base of maybe 200-400 bp.

b) it will also depend on the question you are asking: if you are interested in a particular locus you might have problem if a) applies. if you are interested in general phenomena, who cares if you miss one peak out of 15000?

anyone with real data regardings this issue?

[ramkumar]

Thanks mudshark. A friend has done ChIP-seq with input from C57BL/6 : DBA-2 mixed background and his reference genome has been that of C57BL/6. As you say, globally, there appears to be no problem. but you could be missing peaks.
unless someone has compared the same dataset, from a given genetic background, on two different reference genomes, we can't tell. I also wonder if any genome, other than that of BL/6 has a good coverage.