GWAS-eQTL-Colocalization

GWAS-eQTL Colocalization Analysis Workflow

Zhang Lab@Columbia by Fang Li [2019-08-01]

1. The purpose of GWAS-eQTL intergration

• Is my variant an eQTL?
• Is the leading variant of the GWAS and eQTL signal the same?
• Is my GWAS association of interest driven by an eQTL that may indiciate a functinal mechanism?

GWAS locus that colocalized with eQTL is one of the primary and scalable signal for functional follow-up analyses.

2. Install R/RStudio and packages

• Install the latest version of R or RStudio.
• Install R pakage locuscomparer.
• Install R package coloc:
  if (!requireNamespace("BiocManager", quietly = TRUE)) install.packages("BiocManager")
BiocManager::install("snpStats")
install("coloc")

3. Colocalization analysis using coloc

• Read sample data into R:
  You can download the sample files by right clicking the link: GWAS and eQTL datasets.
  eqtl <- read.table(file="[path to]/Artery_Coronary_v7_eQTL_PHACTR1.txt", header=T, as.is=T); head(eqtl) gwas <- read.table(file="[path to]/CAD_GWAS.txt", header=T, as.is=T); head(gwas)

• Merge gwas and eqtl data sets by only shared “rs _ id”:
  input <- merge(eqtl, gwas, by="rs_id", all=FALSE, suffixes=c("_eqtl","gwas") head(input)

  Optinal: provide suffix to differentiate data source from gwas or eqtl.

• Run coloc using coloc.abf() fuction:

result <- coloc.abf(dataset1=list(pvalues=input$pval_nominal_gwas, type="cc", S=0.33, N=nrow(gwas)) dataset2=list(pvalues=input$pval_nominal_eqtl, type="quant", N=nrow(eqtl)), MAF=input$maf)

Comments: coloc.abf() function needs two named lists (gwas and eqtl) that contain p-values, the type of study(“cc” for case-control studies, “quant” for quantitative traits) and sample size(N). s= the proportion of samples are cases, when type=”cc”. It also needs the minor allele frequency.

• Read out posterior probabilities for colocalization:
  • H0: neither trait has a genetic association in the region
  • H1/H2: only trait 1/trait 2 has a genetic association in the region
  • H3: both traits are associated, but with different causal variants
  • H4: both traits are associated and share a single causal variant

A posterior probability of ≥75% is considered strong evidence of the eQTL-GWAS pair influencing both the expression and GWAS trait at a particular region.

4. Visualization using locuscomparer

• Define file names of the GWAS and eQTL data sets:
  gwas_fn="[path to]/CAD_GWAS.txt"
eqtl_fn="[path to]/Artery_Coronary_v7_eQTL_PHACTR1.txt"
marker_col="rs_id"
pval_col="pval_nominal"

• Run locuscompare to visualize: locuscompare(in_fn1=gwas_fn, in_fn2=eqtl_fn, title1="GWAS", title2="eQTL", marker_col1= marker_col, pval_col1=pval_col, marker_col2=marker_col, pval_col2=pval_col))

• Results output:

• You may download the Rmd file here and the html output here.

Materials here are licensed as CC BY-NC-SA 4.0 Creative Commons License.