SPAGMMATtest: Run single variant or gene- or region-based score tests with...

In weizhouUMICH/SAIGE: Efficiently controlling for case-control imbalance and sample relatedness in single-variant assoc tests (SAIGE) and controlling for sample relatedness in region-based assoc tests in large cohorts and biobanks (SAIGE-GENE)

Run single variant or gene- or region-based score tests with SPA based on the linear/logistic mixed model.

Description

Run single variant or gene- or region-based score tests with SPA based on the linear/logistic mixed model.

Usage

SPAGMMATtest(
  bgenFile = "",
  bgenFileIndex = "",
  vcfFile = "",
  vcfFileIndex = "",
  vcfField = "DS",
  savFile = "",
  savFileIndex = "",
  sampleFile = "",
  idstoExcludeFile = "",
  idstoIncludeFile = "",
  rangestoExcludeFile = "",
  rangestoIncludeFile = "",
  chrom = "",
  start = 1,
  end = 2.5e+08,
  IsDropMissingDosages = FALSE,
  minMAC = 0.5,
  minMAF = 0,
  maxMAFforGroupTest = 0.5,
  minInfo = 0,
  GMMATmodelFile = "",
  varianceRatioFile = "",
  SPAcutoff = 2,
  SAIGEOutputFile = "",
  numLinesOutput = 10000,
  IsSparse = TRUE,
  IsOutputAFinCaseCtrl = FALSE,
  IsOutputHetHomCountsinCaseCtrl = FALSE,
  IsOutputNinCaseCtrl = FALSE,
  IsOutputlogPforSingle = FALSE,
  LOCO = TRUE,
  condition = "",
  sparseSigmaFile = "",
  groupFile = "",
  kernel = "linear.weighted",
  method = "optimal.adj",
  weights.beta.rare = c(1, 25),
  weights.beta.common = c(1, 25),
  weightMAFcutoff = 0.01,
  weightsIncludeinGroupFile = FALSE,
  weights_for_G2_cond = NULL,
  r.corr = 0,
  IsSingleVarinGroupTest = TRUE,
  cateVarRatioMinMACVecExclude = c(0.5, 1.5, 2.5, 3.5, 4.5, 5.5, 10.5, 20.5),
  cateVarRatioMaxMACVecInclude = c(1.5, 2.5, 3.5, 4.5, 5.5, 10.5, 20.5),
  dosageZerodCutoff = 0.2,
  IsOutputPvalueNAinGroupTestforBinary = FALSE,
  IsAccountforCasecontrolImbalanceinGroupTest = TRUE,
  IsOutputBETASEinBurdenTest = FALSE,
  IsOutputMAFinCaseCtrlinGroupTest = FALSE,
  X_PARregion = "60001-2699520,154931044-155270560",
  is_rewrite_XnonPAR_forMales = FALSE,
  sampleFile_male = "",
  method_to_CollapseUltraRare = "absence_or_presence",
  MACCutoff_to_CollapseUltraRare = 10,
  DosageCutoff_for_UltraRarePresence = 0.5
)

Arguments

bgenFile	character. Path to bgen file. Currently version 1.2 with 8 bit compression is supported
bgenFileIndex	character. Path to the .bgi file (index of the bgen file)
vcfFile	character. Path to vcf file
vcfFileIndex	character. Path to index for vcf file by tabix, ".tbi" by "tabix -p vcf file.vcf.gz"
vcfField	character. genotype field in vcf file to use. "DS" for dosages or "GT" for genotypes. By default, "DS".
savFile	character. Path to sav file
savFileIndex	character. Path to index for sav file .s1r
sampleFile	character. Path to the file that contains one column for IDs of samples in the bgen file with NO header
idstoExcludeFile	character. Path to the file containing variant ids to be excluded from the bgen file. The file does not have a header and each line is for a marker ID.
idstoIncludeFile	character. Path to the file containing variant ids to be included from the bgen file. The file does not have a header and each line is for a marker ID.
rangestoExcludeFile	character. Path to the file containing genome regions to be excluded from the bgen file. The file contains three columns for chromosome, start, and end respectively with no header
rangestoIncludeFile	character. Path to the file containing genome regions to be included from the bgen file. The file contains three columns for chromosome, start, and end respectively with no header
chrom	character. string for the chromosome to include from vcf file. Required for vcf file. Note: the string needs to exactly match the chromosome string in the vcf/sav file. For example, "1" does not match "chr1". If LOCO is specified, providing chrom will save computation cost
start	numeric. start genome position to include from vcf file. By default, 1
end	numeric. end genome position to include from vcf file. By default, 250000000
IsDropMissingDosages	logical. whether to drop missing dosages (TRUE) or to mean impute missing dosages (FALSE). By default, FALSE. This option only works for bgen, vcf, and sav input.
minMAC	numeric. Minimum minor allele count of markers to test. By default, 0.5. The higher threshold between minMAC and minMAF will be used
minMAF	numeric. Minimum minor allele frequency of markers to test. By default 0. The higher threshold between minMAC and minMAF will be used
maxMAFforGroupTest	numeric. Maximum minor allele frequency of markers to test in group test. By default 0.5.
minInfo	numeric. Minimum imputation info of markers to test. By default, 0. This option only works for bgen, vcf, and sav input
GMMATmodelFile	character. Path to the input file containing the glmm model, which is output from previous step. Will be used by load()
varianceRatioFile	character. Path to the input file containing the variance ratio, which is output from the previous step
SPAcutoff	by default = 2 (SPA test would be used when p value < 0.05 under the normal approximation)
SAIGEOutputFile	character. Path to the output file containing assoc test results
numLinesOutput	numeric. Number of markers to be output each time. By default, 10000
IsSparse	logical. Whether to exploit the sparsity of the genotype vector for less frequent variants to speed up the SPA tests or not for dichotomous traits. By default, TRUE
IsOutputAFinCaseCtrl	logical. Whether to output allele frequency in cases and controls. By default, FALSE
IsOutputHetHomCountsinCaseCtrl	logical. Whether to output heterozygous and homozygous counts in cases and controls. By default, FALSE. If True, the columns "homN_Allele2_cases", "hetN_Allele2_cases", "homN_Allele2_ctrls", "hetN_Allele2_ctrls" will be output.
IsOutputNinCaseCtrl	logical. Whether to output sample sizes in cases and controls. By default, FALSE
IsOutputlogPforSingle	logical. Whether to output log(Pvalue) for single-variant assoc tests. By default, FALSE. If TRUE, the log(Pvalue) instead of original P values will be output
LOCO	logical. Whether to apply the leave-one-chromosome-out option. By default, TRUE
condition	character. For conditional analysis. Genetic marker ids (chr:pos_ref/alt if sav/vcf dosage input , marker id if bgen input) seperated by comma. e.g.chr3:101651171_C/T,chr3:101651186_G/A, Note that currently conditional analysis is only for bgen,vcf,sav input.
sparseSigmaFile	character. Path to the file containing the sparseSigma from step 1. The suffix of this file is ".mtx".
groupFile	character. Path to the file containing the group information for gene-based tests. Each line is for one gene/set of variants. The first element is for gene/set name. The rest of the line is for variant ids included in this gene/set. For vcf/sav, the genetic marker ids are in the format chr:pos_ref/alt. For bgen, the genetic marker ids should match the ids in the bgen file. Each element in the line is seperated by tab.
kernel	character. For gene-based test. By default, "linear.weighted". More options can be seen in the SKAT library
method	character. method for gene-based test p-values. By default, "optimal.adj". More options can be seen in the SKAT library
weights.beta.rare	vector of numeric. parameters for the beta distribution to weight genetic markers with MAF <= weightMAFcutoff in gene-based tests.By default, "c(1,25)". More options can be seen in the SKAT library
weights.beta.common	vector of numeric. parameters for the beta distribution to weight genetic markers with MAF > weightMAFcutoff in gene-based tests.By default, "c(1,25)". More options can be seen in the SKAT library. NOTE: this argument is not fully developed. currently, weights.beta.common is euqal to weights.beta.rare
weightMAFcutoff	numeric. Between 0 and 0.5. See document above for weights.beta.rare and weights.beta.common. By default, 0.01
weightsIncludeinGroupFile	logical. Whether to specify customized weight for makers in gene- or region-based tests. If TRUE, weights are included in the group file. For vcf/sav, the genetic marker ids and weights are in the format chr:pos_ref/alt;weight. For bgen, the genetic marker ids should match the ids in the bgen filE, e.g. SNPID;weight. Each element in the line is seperated by tab. By default, FALSE
weights_for_G2_cond	vector of float. weights for conditioning markers for gene- or region-based tests. The length equals to the number of conditioning markers, delimited by comma. By default, "c(1,2)"
r.corr	numeric. bewteen 0 and 1. parameters for gene-based tests. By default, 0. More options can be seen in the SKAT library
IsSingleVarinGroupTest	logical. Whether to perform single-variant assoc tests for genetic markers included in the gene-based tests. By default, FALSE
cateVarRatioMinMACVecExclude	vector of float. Lower bound of MAC for MAC categories. The length equals to the number of MAC categories for variance ratio estimation. By default, c(0.5,1.5,2.5,3.5,4.5,5.5,10.5,20.5). If groupFile="", only one variance ratio corresponding to MAC >= 20 is used
cateVarRatioMaxMACVecInclude	vector of float. Higher bound of MAC for MAC categories. The length equals to the number of MAC categories for variance ratio estimation minus 1. By default, c(1.5,2.5,3.5,4.5,5.5,10.5,20.5). If groupFile="", only one variance ratio corresponding to MAC >= 20 is used
dosageZerodCutoff	numeric. In gene- or region-based tests, for each variants with MAC <= 10, dosages <= dosageZerodCutoff with be set to 0. By default, 0.2.
IsOutputPvalueNAinGroupTestforBinary	logical. In gene- or region-based tests for binary traits. if IsOutputPvalueNAinGroupTestforBinary is TRUE, p-values without accounting for case-control imbalance will be output. By default, FALSE
IsAccountforCasecontrolImbalanceinGroupTest	logical. In gene- or region-based tests for binary traits. If IsAccountforCasecontrolImbalanceinGroupTest is TRUE, p-values after accounting for case-control imbalance will be output. By default, TRUE
IsOutputBETASEinBurdenTest	logical. Output effect size (BETA and SE) for burden tests. By default, FALSE
IsOutputMAFinCaseCtrlinGroupTest	logical. Whether to output minor allele frequency in cases and controls in set-based tests By default, FALSE
X_PARregion	character. ranges of (pseudoautosomal) PAR region on chromosome X, which are seperated by comma and in the format start:end. By default: '60001-2699520,154931044-155260560' in the UCSC build hg19. For males, there are two X alleles in the PAR region, so PAR regions are treated the same as autosomes. In the NON-PAR regions (outside the specified PAR regions on chromosome X), for males, there is only one X allele. If is_rewrite_XnonPAR_forMales=TRUE, genotypes/dosages of all variants in the NON-PAR regions on chromosome X will be multiplied by 2.
is_rewrite_XnonPAR_forMales	logical. Whether to rewrite gentoypes or dosages of variants in the NON-PAR regions on chromosome X for males (multiply by 2). By default, FALSE. Note, only use is_rewrite_XnonPAR_forMales=TRUE when the specified VCF or Bgen file only has variants on chromosome X. When is_rewrite_XnonPAR_forMales=TRUE, the program does not check the chromosome value by assuming all variants are on chromosome X
sampleFile_male	character. Path to the file containing one column for IDs of MALE samples in the bgen or vcf file with NO header. Order does not matter
method_to_CollapseUltraRare	character. Method to collpase the ultra rare variants in the set-based association tests. This argument can be 'absence_or_presence', 'sum_geno', or ”. absence_or_presence: For the resulted collpased marker, any individual having DosageCutoff_for_UltraRarePresence <= dosage < 1+DosageCutoff_for_UltraRarePresence for any ultra rare variant has 1 in the genotype vector, having dosage >= 1+DosageCutoff_for_UltraRarePresence for any ultra rare variant has 2 in the genotype vector, otherwise 0. sum_geno: Ultra rare variants with MAC <= MACCutoff_to_CollapseUltraRare will be collpased for set-based tests in the 'sum_geno' way and the resulted collpased marker's genotype equals weighted sum of the genotypes of all ultra rare variants. NOTE: this option sum_geno currently is NOT active. By default, "absence_or_presence".
MACCutoff_to_CollapseUltraRare	numeric. MAC cutoff to collpase the ultra rare variants (<= MACCutoff_to_CollapseUltraRare) in the set-based association tests. By default, 10.
DosageCutoff_for_UltraRarePresence	numeric. Dosage cutoff to determine whether the ultra rare variants are absent or present in the samples. Dosage >= DosageCutoff_for_UltraRarePresence indicates the varaint in present in the sample. 0< DosageCutoff_for_UltraRarePresence <= 2. By default, 0.5.

Value

SAIGEOutputFile

weizhouUMICH/SAIGE documentation built on May 6, 2022, 12:34 a.m.