Signal desert in ChIP-seq. data? (no signal/ noise in some regions)

Question

[chiefcat]

Hi guys,

Do you ever see there are some regions doesn't have any signal (not even noise) in ChIP-seq. data? What is the cause/ meaning of this?? Thanks!

My case is occurred at genic region, like below (Red box):

Answer 1

[Carlo Yague]

There are a least three possibilities :

• That gene regions is lost in your sample due to evolution or genetic manipulation
• That region was masked from your reference genome
• Reads couldn't map to that region (for instance because of low complexity)

[ATpoint]

I would first check if it was masked. If it was low-complexity, one should expect excessive alignments with low MAPQ there, at least in some of the samples, due to the random placement of multimappers by most aligners. And if it was deleted, one should at least see some noisy reads sitting there. A total loss-of-signal is therefore probably a hard-masking in the fasta. You can zoom in on the IGV until you see the nucleotide sequence. If masked, then there should be N's.

[chiefcat]

Hi ATPoint,

Thanks for your detailed explanation and suggestion. Yes, the total loss-of-signal is really due to masking. So, this is just visualization problem that won't affect the peak-calling, right?

[chiefcat]

Hi Carlo,

After checking the nucleotide sequence in IGV as suggested by ATPoint, I found that the missing of the signal is due to masking. Thanks for your input!