Hi nicolas. I had a strange issue this week that I try found the reason. Maybe due to HiC-Pro, maybe not.
-I have a matrix A that I produced at raw resolution of 2kb and 10kb.
-I have a matrix B that I produced at raw resolution of 10kb.
I do not use your iced implementation algorithm (i have my own on my pipeline.
I take A at a resolution of 2 kb, and B at a resolution of 10 kb, and both bin them to a resolution of 100kb. When I make the log2 ratio of A-B, there is 1 bin on the whole genome that is that is not equally represented in the two matrix and that make a huge empty region in the log ratio.
When I take A and B at 10kb before binning, the empty region disipear.
So maybe, there is a small and tricky issue in the read assignation to the bing, specially when the resolution is really small ( resolution < 5 kb).
This sound logical in fact but maybe there is a way to solve it one day.
Have fun with that question, I will investigate on it ;).
Hi nicolas. I had a strange issue this week that I try found the reason. Maybe due to HiC-Pro, maybe not.
-I have a matrix A that I produced at raw resolution of 2kb and 10kb. -I have a matrix B that I produced at raw resolution of 10kb.
I do not use your iced implementation algorithm (i have my own on my pipeline.
I take A at a resolution of 2 kb, and B at a resolution of 10 kb, and both bin them to a resolution of 100kb. When I make the log2 ratio of A-B, there is 1 bin on the whole genome that is that is not equally represented in the two matrix and that make a huge empty region in the log ratio.
When I take A and B at 10kb before binning, the empty region disipear.
So maybe, there is a small and tricky issue in the read assignation to the bing, specially when the resolution is really small ( resolution < 5 kb).
This sound logical in fact but maybe there is a way to solve it one day.
Have fun with that question, I will investigate on it ;).