Skip to content

Heritability vs SNP heritability vs broad-sense vs narrow-sense vs missing heritability

Definition

In GWAS/statistical genetics, these terms answer different questions: heritability is the umbrella concept (genetic contribution to phenotype variance in a population), broad-sense includes all genetic variance components, narrow-sense keeps only additive effects, SNP heritability is the additive variance captured/tagged by measured SNP data and model assumptions, and missing heritability is the gap between higher family-based estimates and lower SNP/discovered-variant estimates.

How they differ

Heritability Broad-sense heritability Narrow-sense heritability SNP heritability Missing heritability
Main idea Umbrella concept: fraction of variance due to genetic differences in a population/environment. Total genetic variance share (additive + dominance + interaction), often denoted H^2. Additive genetic variance share only, often denoted h^2. Portion of variance explained by SNPs measured/tagged in GWAS data and model. The unexplained gap between larger family-based heritability and smaller SNP/discovered-variant estimates.
Typical estimation context Conceptual and methodological framing. Twin/family/experimental designs where non-additive effects can contribute. Family or population methods focused on additive inheritance. GREML, LD score regression, and related summary-statistics methods. Comparison across methods and data layers (family studies vs GWAS outputs).
Relationship to GWAS hits Sets expectations for potential discoverable signal. Usually above SNP-based estimates for complex traits. Closer to what additive GWAS models target. Directly tied to common-variant GWAS architecture and tagging. Highlights what current GWAS/imputation/discovery has not yet captured.

Rule of thumb: For standard GWAS discussions, compare reported SNP heritability to narrow-sense heritability ideas, and treat broad-sense heritability as a wider upper context rather than the direct GWAS target.

Important caveats

  • Heritability is population-, environment-, and phenotype-definition specific; it is not a fixed biological constant for all contexts.
  • A high heritability does not imply determinism at the individual level.
  • Missing heritability does not imply "non-genetic"; it often reflects incomplete tagging, rare/structural variants, architecture complexity, and finite sample size.

References

  • Visscher PM, Hill WG, Wray NR. (2008). Heritability in the genomics era. Nat Rev Genet.
  • Yang J, et al. (2017). Concepts, estimation and interpretation of SNP-based heritability. Nat Genet.
  • Manolio TA, et al. (2009). Finding the missing heritability of complex diseases. Nature.