Perform cell-type analysis using partitioned heritability

A common usage of partitioned heritability is to estimate enrichment and significance for an annotation typical for a specific cell-type. For this analysis, it is useful to adjust for a baseline set of annotations, and then estimate the association for a large set of annotations ("cell-types")

Usage

celltype_analysis(sumstat, covariate_dir, ldscore_dir, weights = NULL)

Arguments

sumstat

A dplyr::tibble() with columns SNP, Z and N

covariate_dir

a directory containing the files ld.parquet and annot.parquet

ldscore_dir

a directory containing the files ld.parquet and annot.parquet. Each annotation will be executed, while adjusting for all annotations in covariate_dir

weights

Optional, a data.frame or tbl with columns SNP, L2

Value

a dplyr::tibble()

Examples

if (FALSE) { # \dontrun{
sumstat1 <- arrow::read_parquet(system.file("extdata", "sumstats.parquet", package = "ldsR"))
sumstat1 <- dplyr::select(sumstat1, SNP, Z = Z.x, N=N.x)


celltype_analysis(
 sumstat1,
 covariate_dir = system.file("extdata", "baseline1.1_test", package = "ldsR"),
 ldscore_dir = system.file("extdata", "superclusters", package = "ldsR")
)
} # }