Input Format¶

This page describes the format of the GWAS summary association data (supplied through the --local-hsqg or --local-rhog flag), genome partition file (supplied through the --partition flag) and the reference panel required by HESS (supplied through the --bfile format).

GWAS summary association data¶

HESS requires a single file in plain text or gzipped text containing the following columns to be present in the GWAS summary association data:

SNP - rs ID of the SNP (e.g. rs62442).
CHR - Chromosome number of the SNP. This should be a number between 1 and 22.
BP - Base pair position of the SNP.
A1 - Effect allele of the SNP. The sign of the Z-score is with respect to this allele.
A2 - The other allele of the SNP.
Z - The Z-score of the SNP.
N - Sample size of the SNP.

For estimating local SNP-heritability, one summary statistics file should be supplied through the --local-hsqg flag. For estimating local genetic covariance, two summary statistics files should be supplied through the --local-rhog flag.

( Note: This file format is compatible with LDSC. However, HESS requires two additional columns than LDSC [CHR and BP]. Therefore, one may not directly run HESS on summary stats in LDSC format unless these two columns are provided. Other columns [e.g. MAF, INFO, etc.] may be included in the file, but will not be used. It is also recommended [although not required] that the summary data files are sorted by their chromosomes and base pairs. All SNPs with either duplicate ID or position will be removed before any analysis. )

( Note: We recommend to run HESS on imputed summary statistics data. HESS may yield downwardly biased estimates when only genotyped SNPs are used. )

Genome partition¶

The genome partition file should be in bed format, one for each chromosome. We recommend to use the approximately LD-independent loci provided by Berisa et al., which can be downloaded here. Please choose the genome partition that corresponds to the GWAS population. The genome partition file is supplied through the --partition flag.

( Note: In order to obtain unbiased estimates, the partitions should be as LD-independent as possible. Using a small window may result in LD leakage and biased estimates. )

Reference panel¶

Reference panels should be in PLINK format (supplied through the --bfile flag).

The following is a list of popular publicly available reference panels.

We provide 1000 Genomes reference panel for Europeans here. All SNPs in this reference panel have minor allele frequency greater than 1%.

( Note: If a SNP has missing genotypes at some individuals in the reference panel, HESS imputes the missing value with the population average. HESS also removes all SNPs with duplicate ID or position. )