HI, all
I have a question when computing region coverage of pair-end RNASeq data.
As showed by the sketch map, when computing region coverage, whether I should use actually mapped reads or extended reads? The coverage for impacted region may be different. I wonder which one is reasonable.
I want to compute and compare coverage for small regions like 100 bp, not as large as a gene. Also, one of the compared sample is sequenced Pair-endly, the other is sequenced single-endly. Which way of computing is more suitable in this case?
I have a Rip-Seq and want to use this RNA-Seq as a control to call peaks. The genome would be binned into 100bp regions and coverage of each region will be computed for both Rip-Seq and RNA-Seq. A following fisher's test would be used to select significantly enriched regions of Rip-Seq.
The region is defined. "How many fragments were sequenced from this region?" is what I want to ask . If one of my defined regions happens to located in the internal region of two ends of a fragment, the coverage of it in RNA-Seq would be 0 if just mapped tags used. However, this region is surely covered by reads.
Thank you very much!
Reference ========--------------------------------------------------===========
Actually mapped reads ^^^^ $$$$
Extended reads ^^^^^^^ $$$$$$$$$$$
Impacted region **** *******
Reference ======================================================================
Actually mapped reads ^^^^ $$$$
Extended reads ^^^^^^^^^^^^^^^^$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
Impacted region *************************************************************
#ignore the relative ratio of ^ and $
= represents exon bases; - represents intron bases
^ represents left-end reads; $ represents right-end reads
Update: Add my usage for this kind of data to make question clear.