I've a list of human genomic regions and I want for each of them a list of spatially (in a 3D genome manner) close genomic regions + associated score. I know some DB using Hi-C data such as :
Hi,
it sounds that the normalized Hi-C contact frequency between two loci of interest would indeed be the measure that you are looking for.
Though, there are at least two caveats. (A) The 3d conformation of a genome varies greatly from nucleus to nucleus, so that two regions that are close to each other in one cell may be far away from each other in another cell. Hi-C maps show the frequency with which two genomic loci come sufficiently close to each other for the method to detect a contact, not the average distance between two loci. The relationship between Hi-C contact frequency and the average distance is still an open question in the field.
(B) Hi-C contact frequencies are greatly determined by the physics of chromosomes and do not necessarily imply biological function. For example, two loci that are located close to each other on the same chromosome would come in contact very frequently, simply because they are connected and cannot really float apart far away. These and other large-scale features of Hi-C map have to be controlled for before drawing any type of biological conclusion.
