Haploview CI

[CPG]

Hi everyone,
I was hoping one of you could help me out with a Haploview question.

The D' value of one of the SNP pairs that I tested was 1, which is consistent with there only being 3 of the 4 possible haplotypes in the data set.

However, the confidence interval upper bound value is 0.99. This means the D' estimate does not lie in the interval. Is haploview not giving proper confidence intervals?
What is the big difference between the confidence limits by Gabriel 2002 and normal confidence intervals?

[chrchang]

The Haploview code has a bit of special-case logic for the D' = 1 posterior probability estimate that may be going wrong here. It shouldn't really distort your results, though.

Have you checked whether PLINK 1.9 --blocks (https://www.cog-genomics.org/plink2/ld#blocks ) gives you the correct confidence interval here?

[CPG]

Thank you for the suggestion.
So this doesn't have anything to do with Gabriel confidence limits not being proper confidence intervals?

[chrchang]

Correct, the basic Gabriel CI definition is valid.