select_mamba_loci: Select loci to include in mamba() model, combining clumped...

In dan11mcguire/mamba: MAMBA: Meta-analysis Model Based Assessment of Replicability

Description

Given 1. a set of clumped significant SNPs from meta-analysis which we want to assess with MAMBA, and 2. a random set of pruned markers in approximate LD, Identify pruned variants no closer than (bpdiff) away from clumped variants, to be used in fitting mamba, to be used in fitting mamba() model. The inclusion of randomly pruned variants allows the non-replicable zero-effect cluster of the mixture model to be reliably estimated.

Usage

select_mamba_loci(meta_file, clump_file, bpdiff = 5 * 10^5L,
  prunedsnps_file = NULL, pthresh = NULL)

Arguments

meta_file	meta-analysis association file (with columns as outputted by mamba::make_mamba_summary_files()$assoc. i.e. plink assoc file format )
clump_file	.clumped file output from plink identifying sentinel variants of interest for mamba model
bpdiff	BP distance between a randomly pruned snp and a clumped snp of interest.
prunedsnps_file	a file with a set of SNP-ids (1 per line) which are approximately independent according to a reference panel. If NULL, a set of SNP id's (in chr:bp_ref_alt format) from HRC reference panel are used. These were created using plink with parameters –indep-pairwise 500kb 1 0.1, –maf 0.01
pthresh	maximum p-value for index snps of interest. If not provided, this is taken as maximum p-value of the clumped sentinel variants.

Value

a data.table with the inputted meta_file, subset to SNPs which are intended to be included in the mamba model. The columns index_snp and prunesnp are indicators for whether the SNP belonged to either the clumped_file or is a randomly pruned marker.

Examples

 ## Not run: 
   ## see online vignette for detailed example. 
    

## End(Not run)

dan11mcguire/mamba documentation built on Nov. 10, 2020, 12:37 a.m.