What is the structure of the data in Haplotype reference panel?

Question

[kirannbishwa01]

I am trying to work on haplotype phasing and would like to see how the data is represented in "haplotype reference panel".

Is "haplotype reference panel" a vcf file like this: http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/release/20130502/ or is it something else.

Several papers talk about "reference haplotype" from several samples. So, is there a universal method to represent "reference panel"? How are other informations (recombination etc.) embedded in these reference panels?

Thanks,

[Kevin Blighe]

Just for everyone else: this is a continuation of the discussion Custom Reference panel creation for data imputation from .vcf files

Are you asking about the format of these data sructures just out of curiosity? I believe that they are structured differently depending on the program using them, but the meaning behind the data is always the same.

As mentioned, a reference panel is just a list of variants, like in VCF format. Each program will then most likely encode these variants in its own format, like GEN format for the IMPUTE2 program. If you look HERE, you'll see this.

A reference haplotype, then, is an abstract entity that's derived from the reference panel. A reference haplotype would be produced by looking at linkage disequilibrium statistics across the reference panel, and then effectively breaking this panel up in the aforementioned entities (haploblocks). The testing / sample data is then compared against these haploblocks and imputations are made. At the very top of the page to which I linked you (above), you'll see a very basic illustration of this, where a particular haploblock is given as 15 SNPs.

Not sure if this helps to clarify.

[kirannbishwa01]

Thank you Kevin. I asked it out of curiosity as of now; but will be working on "haplotype phasing" project of my own soon.

Yes, that picture helps. Thanks a lot.