Average Width Of A Heterochromatin Block In Mouse/Human?

Question

[nbartonicek]

I am trying to model heterochromatin with epigentic marks (H3K9me3 and H3K27me3). Does anyone know what is the average size of heterochromatin blocks in mammals? Or the size distribution, for that matter.

Any information useful!

Cheers,

Nenad

Answer 1

[Ben]

I don't know if this has already been published, but if not a size distribution could easily be drawn up from Bing Ren's human and mouse Hi-C "topological domains" dataset, as called by their "directionality index" method (data). Else for better hetero/euchromatin distinction, the PC eigenvectors from Lieberman-Aiden et al. (2009) could be used in the same way with a bit more fiddling about.

Answer 2

[PoGibas]

You should check UCSC Broad ChromHMM track.
Chromatin is divided into 15 states (#13 is heterochromatin, #12 is Polycomb etc.). You can extract chromatin state of interest, it's start & end coordinates (thickStart & thickEnd) genome wide and calculate size distribution by yourself.

Edit
This recent ENCODE paper (Integrative annotation of chromatin elements from ENCODE data) nicely reviews chromatin states, compares ChromHMM & Segway annotations. Links to the data in the article.

[Ben]

Yeah that's an easier way, though there might be some circular reasoning down the line depending on what the OP is doing (because the HMM inputs were histone marks and other chromatin features).