Understanding the genetic architecture of complex human phenotypes is a fundamental question to the field of genetics, with broad implications for identifying genes related to disease and predicting individual risk profiles1-6. A central element of this problem is estimating narrow-sense heritability (h2), the fraction of phenotypic variation in a population determined by genetic variation under an additive model7. While the last decade of genome-wide association studies (GWAS) produced thousands of novel loci associated with hundreds of phenotypes8, the sum of their effects ( hgwas2) explain only a small fraction of the estimated heritability for most phenotypes5. The gap between hgwas2 and h2 is called the “missing heritability” and several explanations for this difference have been posited, including upward bias in estimates of h22,4,9. The objective of this work is to develop a method for estimating h2 (defined in Methods) that (1) does not require closely related individuals, (2) can be applied to both quantitative and case-control phenotypes, and (3) is able to localize narrow-sense heritability to individual chromosomes or other genomic segments.
