I am not understanding the concept of haplotyping in GWAS. I think that we can only make haplotypes in case where we have dense SNPs throughout the genome which are completely in LD or we can make haplotype when we have very less number of SNPs which might be in LD or not. Or haplotyping denotes some other thing in case of GWAS? Thanks,
A haplotype can be one base pair or span an entire chromosome. With population data you can phase the heterozygous SNPs/SNVs to tell you which base goes with which strand. With low density SNP arrays you can still identify stretches of mutations that are always found on the same strand / haplotype.
One consideration is resolution. Lets say you have two phased hets (A:G, T:C) 3 Mb apart. Just because you see A(pos 1) and T(pos 2) together all the time does not guarantee that they are on the same haplotype. There could be a lot of mutations in between that break up the haplotype. The denser the genotyping the better haplotype predictions.
Thanks for the suggestion.
Here is an Annual Review article on the medical importance of haplotypes: