Chromatin States
Chromatin states are a genome annotation. The idea behind these is that different combinations of histone marks (and
potentially other chromatin features) are indicative of the function of genomic regions. For example, chromatin marked
by H3K4me3 and H3K27ac is usually considered to be an active promoter. There are several tools that use ChIP-seq data
data as input and assign a 'chromatin state' to every region in the genome thereby creating the annotation. Currently, the most
well known of these tools is ChromHMM, but others (e. g. Seqway, EpiCSeg) are gaining in popularity. For a slightly longer
introduction to functional chromatin annotations,
here
is a link to a Bitesize Bioinformatics video.
Like all genome annotations, chromatin states should be treated with care. The segmentation is influenced by many factors,
for example how many marks were used, how good the underlying ChIP data was, which tool was used, the tile size, how many
state were created and so on. Consider that the underlying data could have been interpreted very differently and be aware that
annotations with the same name don't necessarily mean the same thing (take 'poised enhancers' as an infamous example). Also
make sure that the annotation matches your cell type (and organism!) - chromatin landscapes are very different between cell types! For in
depth analysis use the actual ChIP data.
The Chromatin State Explorer uses published functional genome annoations. We have no influence on these nor have we done
any quality control on them. Please cite the original publications.
What the Chromatin State Explorer does
When you upload a set of regions (e. g. ChIP-seq peak coordinates), the app will take the centre point and
determine the state annotation for the selected coordinates. It will then calculate the percentage of regions
falling into each of the states.
Chromatin states are not evenly distributed across the genome. For example, most of the genome is usually
labelled 'background' whereas bivalent regions occur infrequently. Thus, the app also calculates an enrichment for
the regions of interest over the genome representation of each state.
Just as a cautionary note: bivalent regions can be so infrequent that they easily come up as enriched. It's always
worth looking at the original numbers.
Using the Chromatin State Explorer
To view into which chromatin states your regions of interest fall, you need to specify the annotation you want to use,
upload a file containing the regions and specify the file format. Make sure your coordinates are based on either GRCh38 or
GRCm38/mm10.
The Chromatin State Explorer can directly use Seqmonk probe reports as input or a more generic format with the columns
Chromosome, Start and End either as a tab delimited text file with a header, or a comma separated file without a header (e. g.
saved from Excel).
Note that the Chromatin State explorer does not perform a check if the provided genomic coordinates are all present in
the annotation. It will summarise any unfound regions in a new state called 'not found'. There is a checkbox to exclude this artificial
state, but this should be used with caution: If there are a lot of regions that didn't get assigned a state, it is very
likely that something has gone wrong (e. g. wrong organism/genome assembly).
All Chromatin State Explorer annotations use GRCh38 or GRCm38/mm10 coordinates. If the original annotation was for a different
genome build, it has been lifted over.
Development
The Chromatin State Explorer is under development. Please get in touch with any bug reports, questions or requests
for additional annotation sets/functionality (christel.krueger@babraham.ac.uk).