Protocols

Below we provide step-by-step scripts from three papers. They are designed to be run as bash scripts on a UNIX operating system, but should also work in Terminal on a MAC (but are not suitable for Windows). A backslash (\) at the end of a line indicates a command continues onto the next line. Where mentioned, commands can be run in parallel; access to a cluster is highly recommended when performing imputation and when calculating SNP weightings for data which includes rare variants.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Reevaluation of SNP heritability in complex human trait (published in Nature Genetics)

The first three files provide step-by-step instructions for estimating SNP heritability using REML, starting with raw genotype data. The imputation guide requires a (phased) reference panel; we used 1000 Genomes Phase 3 (October 2014), which we downloaded from the IMPUTE2 website. The fourth file provides tips for advanced analyses, including computing enrichment of a SNP set and estimating heritability when there are rare variants.

Imputation of Cohorts
Merging Cohorts
Estimating SNP Heritability

Advanced Analysis

Weblinks:
PLINK | SHAPEIT | IMPUTE2 | Genome Browser
LiftOver which uses the files RsMergeArch.bcp.gz and SNPHistory.bcp.gz
Details of our imputation regions
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Exposing flaws in S-LDSC; reply to Gazal et al. (currently on Bioarxiv))

In this short paper, we explain why correlations between LD scores and test statistics are to be expected under the LDAK Model, and provide a simple demonstration that S-LDSC over-estimates enrichments of heritability.

Exposing Flaws
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Better estimation of SNP heritability from summary statistics provides a new
understanding of the genetic architecture of complex traits (currently on Bioarxiv)

The scripts below provide step-by-step details for each of the analyses we performed. For a fuller explanation, see the relevant sections in SumHer.

Obtain reference panel, format summary statistics, estimate SNP heritability and confounding
Estimate enrichments and genetic correlations
Construct polygenic risk scores and simulate phenotypes
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

If you spot any errors in the guides, please let me know!