APEX: variance-covariance / LD storage guide

This page describes how to store variance-covariance matrices, which capture covariate-adjusted linkage disequilibrium (LD), in APEX. These files are required for multiple-variant analysis from summary statistics (without individual-level data), including meta-analysis. Once installed, you can quickly get started by running ./apex store --help.

Overview

We recommend running apex store separately on each chromosome, which can be accomplished by specifying --region chr1. Each chromosomal vcov file is indexed by chromosomal position and byte offset, allowing fast access to variance-covariance data within a specific region or a specific variant. See here for information on expected vcov file sizes.

apex store should be run with the same set of input files, output prefix, and other options used to store association summary statistics with apex cis. This ensures that both vcov and sumstat files include the same set of individuals, and the same set of covariates, and that apex meta and apexR are able to easily link sumstat and vcov files.

Note that APEX vcov files capture covariated-adjusted LD (the covariance of genotype residuals), which can differ substantially from unadjusted LD in structured samples. Results from apex meta and the apexR R package are designed to be (nearly) numerically equivalent to results from individual-level data – this requires covariate-adjusted (rather than raw) LD. For detailed descriptions of input file formats, please see the input file documentation page.

Table of Contents

1. Running APEX Store
2. Command line arguments

Return to APEX main page.

Running APEX Store

Example command:
./apex store --vcf {vcf} --bed {trait-file} --cov {covariate-file} --prefix {out-name}

Output files. The above command generates 3 output files, {out-name}.vcov.bin, {out-name}.vcov.idx.gz, {out-name}.cis_long_table.tsv.gz. These files store LD data, LD index and covariate adjustment terms, and tabix index respectively. In general, these files are not intended to be human-readable, but can be queried and used for multiple-variant analysis using apex meta and apexR.

Command line arguments

A partial list of options is given below. Please run ./apex store --help to see a complete list of command line flags and options.

• General options
  • --window {BP}, -w {BP} : Window size in base pairs for cis-xQTL analysis. LD will be stored within twice the specified window size.
• Computational resources
  • --threads {N} : No. threads to be used (not to exceed no. available cores).
  • --low-mem : Reduce memory usage by reading and processing genotypes in chunks.