Odd MACS Warning

[jrmmhughes]

Hi

Has anybody seen this error or understand its implications?

WARNING @ Thu, 17 Jun 2010 16:37:31: Treatment tags and Control tags are uneven! FDR may be wrong!


simple single Illumina chipseq experiement. Comparing treatment against the chip input, both datasets are 36bp reads

Cheers

Jim

[ffinkernagel]

this means that the number of aligned tags in treatment and control are 'very' different, and the fdr is going to be even more of a guestimation than it normally is.

I believe, 'very' means more than factor 2, but I might remember incorrectly.

[jrmmhughes]

Ah Thanks Ffinkernagel

That makes sense.


Cheers

Jim

[pkstarstorm05]

This an old thread, but following up on that to anyone out there - if I do have uneven ChIP and input tags, what do I do about it? Is it necessary to do anything about this? Will it throw off downstream analyses, etc?

Cheers,

Paul