search

hiclib / pastis

Poisson-based algorithm for stable inference of DNA Structure
http://hiclib.github.io/pastis/
Other
34 stars 15 forks source link

Simulating HiC data #47

Closed tarak77 closed 5 years ago

tarak77 commented 5 years ago

Hi Nelle,

Thanks for the HiC data simulation code from the 3D coordinates. The example script works well on the given yeast genome data, I also reconstructed a contact matrix from the output coordinates obtained from a mouse single cell data (Setting alpha=-3 and a large beta so as to avoid generating very sparse data)

Will the reconstructed contact matrix in this case be comparable with the original contact matrix? Th parameter alpha=-3 is obtained by relating distances with the contact count, but since single cell data does not have contact frequencies, I wonder how relevant the alpha parameter is? I am finding it difficult to convince myself.

Any help will be great!

Thanks, Tarak

NelleV commented 5 years ago

Sorry… i totally missed this ticket. It's hard to say: the data you are using probably doesn't follow a poisson distribution, so you might be better using a whole other way to simulate data than the one I used.