Approved research

Components of heritability in a UK Biobank cohort

Harvard School of Public Health

Lay summary

We will analyze heritability of several polygenic traits. We will use existing methods and methods under development for partitioning heritability by functional annotation (e.g. cell-type-specific enhancer regions, gene pathways, etc.) to learn about underlying trait biology. We will also examine how SNP heritability varies across LD and MAF categories. Finally, we will evaluate missing heritability using new methods to estimate heritability explained by haplotypes, narrow-sense heritability (using PSMC) and epistatic components of heritability (using Hadamard products). We plan to study a wide range of health-related phenotypes, including diseases and quantitative traits like height and BMI. The data in the UK Biobank?s cohort will allow us to partition heritability at higher resolution and to evaluate missing heritability. These will inform both our understanding of trait biology and the design of future genetic studies. Both of these outcomes will benefit attempts to find actionable drug targets for human disease. Moreover, the methods we develop for partitioning heritability will be published and made open-source for use by the broader research community. We will work with annotations from Finucane et al. 2015 Nat Gen as well as gene sets and new annotations from the ENCODE and Roadmap Epigenomics Consortia and others. We will apply LD score regression [Finucane et al. 2015 Nat Genet], BOLT-REML [Loh et al. 2015 Nat Genet], and a new method under development to assess heritability enrichment of these annotations, as well as enrichment/depletion by LD and MAF, within/across traits and populations. We will also apply new methods to estimate heritability explained by haplotypes, total narrow-sense heritability and epistatic components of heritability. We will analyze the full cohort.

Scope extension:

(Current)

(New)

We will analyze heritability of several polygenic traits. We will use existing methods and methods under development for partitioning heritability by functional annotation (e.g. cell-type-specific enhancer regions, gene pathways, etc.) to learn about underlying trait biology. We will also examine how SNP heritability varies across LD and MAF categories. Finally, we will evaluate missing heritability using new methods to estimate heritability explained by haplotypes, narrow-sense heritability (using PSMC) and epistatic components of heritability (using Hadamard products). We plan to study a wide range of health-related phenotypes, including diseases and quantitative traits like height and BMI. We will work with annotations from Finucane et al. 2015 Nat Gen as well as gene sets and new annotations from the ENCODE and Roadmap Epigenomics Consortia and others. We will apply LD score regression [Finucane et al. 2015 Nat Genet], BOLT-REML [Loh et al. 2015 Nat Genet] and new methods under development to assess heritability enrichment of these annotations, as well as enrichment/depletion by LD and MAF, within/across traits and populations; in particular, we will assess the contribution of functional annotations that are enriched for heritability to mean trait differences between populations. We will also apply new methods to estimate heritability explained by haplotypes, total narrow-sense heritability and epistatic components of heritability. As in our previous work (e.g. Gusev et al. 2014 Am J Hum Genet), we will use applications of genetic risk prediction to validate our estimates of components of heritability. We will also quantify heritability at the local level, e.g. the heritability associated to individual genes.