"Good" epigenetic histone modification ChIP-seq datasets

Question

[Bohdan Khomtchouk]

Could anyone recommend any "good" publicly available (e.g., from the SRA/GEO) ChIP-seq datasets with broad histone marks? In general, I'd like to use these data for testing out novel (and existing) ChIP-seq methodologies for benchmarking purposes.

I'm curious to see if the community considers any specific existing epigenetic histone modification ChIP-seq datasets to be particularly "clean" (e.g., high SNR ratio of peaks relative to noise in gene deserts) that would be particularly suitable as training data, for example, for development and testing of new algorithms. Note I am not interested in transcription-factor ChIP-seq data, only histone modification ChIP-seq data (it does not matter what the specific histone mark under study is, as long as the data is suitable for the purposes outlined above).

Relevant post: Chip-Seq Datasets For Cancer Genome Samples?

[Giovanni M Dall'Olio]

What about the ENCODE data?

Answer 1

[Emily]

ENCODE, Roadmap Epigenomics and Blueprint. You'll find a few thousand ChIP-seq experiments between those three.

Answer 2

[cbio]

Here's a dataset generated by my lab for a paper. The Chip-Seq data is solid, hopefully it'll be helpful for what you're trying to accomplish.

Answer 3

[John]

IHECis the next-generation international Epigeneomic sequencing initiative, although not all the data has been sequenced/released yet.

Personally, I would look at getting some of Germany's contribution to IHEC (data from DEEP) as one of our goals is to produce/publish extremely deep sequencing for histone modifications, RNA, WGBS, DNAse, etc. For a training dataset I don't think you can get much better :)

Answer 4

[jotan]

All the databases suggested above are good places to download.

The commonly studies histone modifications tend to have the cleanest signals.

The H3K4me3 antibodies are very good and almost all the K4me3 ChIP-seq datasets are clean and crystal clear.

H3K36me3 is also tends to have yield good ChIP-seq data but the read counts must be sufficiently high. Any recently generated data should be suitable.

The Jenuwein group has also published some decent H3K9me3 ChIP-seq in mouse cells