CoPheScan: Input data
In cophescan: Adaptation of the Coloc Method for PheWAS

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "Input-"
)

The input dataset for a trait (querytrait) should contain the summary data for SNPs in a genomic region around the query variant (querysnpid) and should have the following fields:

For a Case-control dataset

beta: $\beta$ or effect size

varbeta: variance of $\beta$ or square of the standard error of $\beta$

snp: SNP identifier which maybe rsid or CHR_BP_REF_ALT or CHR_BP

type:'cc'

N: sample size

For a Quantitave dataset

When, beta and varbeta are not available the following

beta: $\beta$ or effect size

varbeta: variance of $\beta$ or square of the standard error of $\beta$

snp: SNP identifier which maybe rsid or CHR_BP_REF_ALT or CHR_BP

type:'quant'

N: sample size

sdY: for a quantitative trait, the population standard deviation of the trait.

Additional fields in case of missing beta/varbeta or sdY

MAF: Minor allele frequency (only required when either beta/varbeta or sdY are unavailable)

pvalues: only required when beta/varbeta are unavailable

s: fraction of samples that are cases (only for a case-control trait when beta/varbeta are unavailable)

library(cophescan)

Explore the data structure of the example dataset available in the cophescan package

data("cophe_multi_trait_data")
trait_dat = cophe_multi_trait_data$summ_stat$Trait_1
str(trait_dat)

Additional field for cophe.susie

LD: Linkage Disequilibrium matrix with row and column names being the same as the snp field.

trait_dat$LD = cophe_multi_trait_data$LD
str(trait_dat$LD[1:10, 1:10])

It is important to check that there is alignment of alleles for which the beta is reported and those in the LD matrix. This can be verified either using coloc::check_alignment or performing a diagnostic check using the susie package https://stephenslab.github.io/susieR/articles/susierss_diagnostic.html.

Note

The input summary data structure for a trait is the same format as required by coloc. To gain a detailed understanding of the input please see the coloc vignette https://chr1swallace.github.io/coloc/articles/a02_data.html and the test datasets provided with the coloc package.
More information on the input can be obtained from the coloc documentation : ?coloc::check_dataset