Learn R Programming
GRAB (version 0.2.4)

GRAB.Region: Perform region-based association tests

Description

Tests for association between phenotypes and genomic regions containing multiple genetic variants, primarily low-frequency and rare variants.

Usage

GRAB.Region(
  objNull,
  GenoFile,
  OutputFile,
  GenoFileIndex = NULL,
  OutputFileIndex = NULL,
  GroupFile,
  SparseGRMFile = NULL,
  MaxMAFVec = "0.01,0.001,0.0005",
  annoVec = "lof,lof:missense,lof:missense:synonymous",
  control = NULL
)

Value

The function returns NULL invisibly. Results are saved to four files:

  1. OutputFile: Region-based test results (SKAT-O, SKAT, Burden p-values).

  2. paste0(OutputFile, ".markerInfo"): Marker-level results for rare variants (MAC >= min_mac_region) included in region tests.

  3. paste0(OutputFile, ".otherMarkerInfo"): Information for excluded markers (ultra-rare variants or failed QC).

  4. paste0(OutputFile, ".infoBurdenNoWeight"): Summary statistics for burden tests without weights.

For method-specific examples and output columns and format, see:

  • POLMM method: GRAB.POLMM.Region

Arguments

objNull

(S3 object) Null model object from GRAB.NullModel. Currently supports POLMM_NULL_Model.

GenoFile

(character) Path to genotype file (PLINK or BGEN format). See GRAB.ReadGeno for details.

OutputFile

(character) Path for saving region-based association results.

GenoFileIndex

(character or NULL) Index files for the genotype file. If NULL (default), uses same prefix as GenoFile. See GRAB.ReadGeno for details.

OutputFileIndex

(character or NULL) Path for progress tracking file. If NULL (default), uses paste0(OutputFile, ".index").

GroupFile

(character) Path to region definition file specifying region-marker mappings and annotation information. Tab-separated format with 2-3 columns per region.

SparseGRMFile

(character or NULL) Path to sparse GRM file (optional).

MaxMAFVec

(character) Comma-separated MAF cutoffs for including variants in analysis (default: "0.01,0.001,0.0005").

annoVec

(character) Comma-separated annotation groups for analysis (default: "lof,lof:missense,lof:missense:synonymous").

control

(list or NULL) List of the following parameters:

  • impute_method (character): Method for imputing missing genotypes: "mean", "minor", or "drop". Default: "minor".

  • missing_cutoff (numeric): Exclude markers with missing rate > this value. Range: 0 to 0.5. Default: 0.15.

  • min_mac_region (numeric): Minimum MAC threshold; markers with MAC < this value are treated as ultra-rare variants. Default: 5.

  • max_markers_region (integer): Maximum number of markers allowed per region. Default: 100.

  • r.corr (numeric vector): Rho parameters for SKAT-O test. Range: 0 to 1. Default: c(0, 0.1^2, 0.2^2, 0.3^2, 0.4^2, 0.5^2, 0.5, 1).

  • weights.beta (numeric vector): Beta distribution parameters for variant weights (length 2). Default: c(1, 25).

  • omp_num_threads (integer): Number of OpenMP threads for parallel computation. Default: data.table::getDTthreads().

  • min_nMarker (integer): Minimum number of markers required for region analysis. Default: 3.

  • SPA_Cutoff (numeric): Z-score cutoff for saddlepoint approximation. When the absolute value of the test statistic exceeds this cutoff, SPA is used to calculate more accurate p-values. Default: 2.