Discordant Haplotype Power Calculator

Are you planning a discordant haplotype (DH) sequencing project?

You can use this tool to help design the study: calculate the sample size, power, or the allowable difference from perfect risk haplotype enrichment.

A "risk haplotype" is a multi-locus haplotype that best separates cases from controls in an association study. A risk haplotype is likely to be the backbone upon which functional sequence variant/s (FSV) lie.  A homozygous DH sequencing design is one that compares:

  • Cases selected to be homozygous for the risk haplotypes; and
  • Controls selected to have zero copies of the risk haplotype.

The power to identify the underlying FSV/s using a DH sequencing design depends on how effectively the risk haplotype backbone has enriched for the FSV/s. If the risk haplotype is a perfect proxy for the FSV/s, then there would be perfect enrichment such that the FSV lies on every risk haplotype and never on anything else. In this perfect (and unlikely) scenario, the resulting FSV frequency would be 1.0 in cases (risk haplotype homozygous) and 0.0 in controls (no risk haplotype). More reasonable assumptions are that the FSV will be enriched, but remain different (δ) from perfect enrichment. If a researcher will accept as possible candidate FSV/s those sequence variants (SVs) that are observed with at least a frequency 1-δ in cases, and at most δ in controls, then δ is the difference from perfect enrichment.

Parameters used by the calculator are:

  • Significance-level (α) to assess SV frequency differences between cases and controls (5×10-8 is standard)
  • Power (1-β) the study is aiming for (0.80 is standard)
  • Difference from perfect risk haplotype enrichment (δ) for determining candidate FSVs (δ=0.1 suggested; indicating a definition of candidate FSV if the SV is found at a frequency≥0.9 in cases and ≤0.1 in controls)
  • Sample size (n)

The first parameter is required. For the remaining three, the user specifies two and the discordant haplotype calculator solves for the third.