Is Haploview The Best Method For Defining Ld Blocks?

Question

[User 7433]

Hi all,

I am working with haploview to view my data and get an idea on how much LD there is across my gene.

Haploview gives three options for defining blocks of LD 1) confidence intervals -95% confidence bounds on D prime are generated and each comparison is called "strong LD", "inconclusive" or "strong recombination". 2) The Four Gamete Rule which detects where 4 haplotypes are observed and thus recombination. 3)Solid Spine of LD - a "spine" of strong LD running from one marker to another along the legs of the triangle in the LD chart (this would mean that the first and last markers in a block are in strong LD with all intermediate markers but that the intermediate markers are not necessarily in LD with each other).

I am putting this data in my thesis and was wondering if anyone knows what is the best method for defining the blocks? Each of these three options can be twiddled so that the settings are more of less stringent...but I am not sure which one is most informative, or most respected.

Can anyone with experience help or point me in the direction of the best method? Also I have looked through the manual..there isn't any way to show all pairwise values in the LD plot...not sure if anyone knows if you can extract this?

Thank you!!!

Answer 1

[Larry_Parnell]

I would look at the methods used by Gil McVean and others in determining recombination hotspots across the human genome (data links: here and here). I would also read papers by DW Bowden on MYH9 (1 example) as I know that they narrowed the susceptibility region (end-stage renal disease) with LD and recombination hotspots.

It may be that all three methods are more or less equally informative for your region, thereby giving "fuzzy" boundaries. Recall that LD boundaries are not as precise as a single SNP coordinate - although results from Haploview, HelixTree and others often give that impression.

[Khader Shameer]

Larry, please add link to the papers you discussed. Thanks !

[Larry_Parnell]

Done! Some are to papers, others to data.

[Khader Shameer]

Thanks Larry, appreciate that.