Haplotype block in PLINK using r2 values

Question

[GabrielMontenegro]

I am trying to compute haplotype blocks using PLINK. I found this function very useful --blocks, as it outputs the haplotypes (with the corresponding SNPs within it) in rows and not as a table (as when using --r2). However, it uses D-prime to estimate LD between SNPs.

Is it possible to get haplotype blocks defined by say SNPs with r2 > 0.8 in the same format i.e. haplotypes by rows with the list of SNPs within each haplotype, and not a pair-wise LD table?

Thanks.

[Kevin Blighe]

If the PLINK developer does not respond, you may consider exporting to Haploview and deriving the haploblocks and types there.

[GabrielMontenegro]

--blocks is the implementation of Haploview!

[Kevin Blighe]

In Haploview, I have derived haploblocks with r2 in the past, though.

[GabrielMontenegro]

Oh, I didn't know that thanks. I want to do this genome-wide and I have a large num of SNPs. Not sure if haploview will be able to handle that

[Kevin Blighe]

Indeed, Haploview would crash with too much data...!