Can HiC-Pro used to map capture Hi-C data ?

[Takeshi]

Hi Nicolas, 

I want to know HiC-Pro can be used to map capture Hi-C data.

Capture Hi-C detects only pair-end reads of interest, so mapped data can include some meaningful biases (uneven coverage). 

So, I wonder if HiC-Pro can be used even in that situation.

Sincerely,

Takeshi

[nservant]

Hi Takeshi,

Thank for your question.
Yes, in practice I already used HiC-Pro to analyse capture Hi-C data.

The only difference with a classical Hi-C experiment is that you do not need to build the map.
So you can use HiC-Pro in stepwise mode with "-s mapping -s proc_hic -s quality_checks -s merge_persample"

Then I added the makeViewpoint script (in bin/utils), which uses the list of valid interactions + the list of restriction fragment + the capture region (as a BED file) and which generates a bedgraph profile, similar to a 4C view.
I used it for one project and it gave good results. If you try it and have any idea of improvments, I would be happy to add them.
Best wishes
Nicolas

[Takeshi]

Hi Nicolas,

Thank you for explanation. I understand.

Can the "makeViewpoint" script be used to generate a bedgraph file
by using (1) validPairs files obtained from a standard Hi-C data, and (2) a bed file for fragments of interest (such as promoters&enhancers) ?
I want to know the script can be used for non-capture Hi-C data in principle.

> The only difference with a classical Hi-C experiment is that you do not need to build the map.

And, I also want to build the corresponding contact map.
Can I make the map in the usual way of HiC-Pro ?
( Is it appropriate to normalize the raw contact map by ICE for capture Hi-C data ? )


Sorry to trouble you,

Takeshi




2016年7月13日水曜日 22時16分29秒 UTC+9 nservant:

[nservant]

Yes, the makeViewpoint script should also be used for standard Hi-C with a simple BED.
And It will generate a viewpoint for each feature of the BED file.
However, my feeling is that it migth be a little bit slow ... but this is something I can look at.

I'm not sure to see the interest of generating the maps if you just capture a region of interest in the genome.
But maybe we are taking about promoters capture Hi-C ?

If so, I never test HiC-Pro in that context ...
My first guess is that the detection of valid 3C products should work.
Then, I think that you will need a dedicated software such the CHICAGO package (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4908757/)
Hope it helps
Nicolas

[Takeshi]

Hi Nicolas,

Thank you for explanation.

> I'm not sure to see the interest of generating the maps if you just capture a region of interest in the genome.
> But maybe we are taking about promoters capture Hi-C ?
> If so, I never test HiC-Pro in that context ...
> My first guess is that the detection of valid 3C products should work.
> Then, I think that you will need a dedicated software such the CHICAGO package (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4908757/)

I understand that
I need a BED file designed for capturing promoter-enhancer contacts
in order to analyze promoter capture Hi-C data (for example) by using CHICAGO.
Is it correct ?

*********************

I'm interested in generating (1) contact profiles of promoter-enhancer or something like that
and (2) the full contact map for data-driven modelling of 3D chromosomes,
from a published (non-promoter) capture Hi-C data.

Actually, I tried to construct a contact map from the published capture Hi-C data by using HiC-Pro,
and then obtained a map quite similar to that from a standard Hi-C data.
But the mapped data can generally include some meaningful biases (uneven coverage),
so I don't know whether use of ICE normalization (for removing uneven coverage) is appropriate or not. 

Best,

Takeshi



2016年7月14日木曜日 6時58分15秒 UTC+9 nservant:

[nservant]

Hi Takeshi

I've never used CHICAGO myself. I guess you will need a BED file to describe your set of promoters.
If I'm correct, CHICAGO will need at some point the list of valid 3C products.
You can get this list of valid 3C products using HiC-Pro. You can also build the map if it is of interest for you.
In addition to the list of the 3C products, the makeViewpoints utils should help you in generating 4C-like profiles for each promoter.
Regarding the normalization, I do not know what people are usally using for promoter capture C data, but I'm not sure that the ICE normalization is the best way to go ...
Maybe you can contact the CHICAGO's authors to get their feedbacks ?
Best whishes
Nicolas