qtlFinder: Distance-based signal identification

In gap: Genetic Analysis Package

Distance-based signal identification

Description

Distance-based signal identification

Usage

qtlFinder(
  d,
  Chromosome = "Chromosome",
  Position = "Position",
  MarkerName = "MarkerName",
  Allele1 = "Allele1",
  Allele2 = "Allele2",
  EAF = "Freq1",
  Effect = "Effect",
  StdErr = "StdErr",
  log10P = "log10P",
  N = "N",
  radius = 1e+06,
  collapse.hla = TRUE,
  build = "hg19"
)

Arguments

d	Input data frame containing summary association statistics.
Chromosome	Chromosome column.
Position	Genomic position column (bp).
MarkerName	Variant identifier (e.g. rsID or SNP ID).
Allele1	Effect allele.
Allele2	Other allele.
EAF	Effect allele frequency.
Effect	Effect estimate.
StdErr	Standard error of the effect estimate.
log10P	Column containing log10(P) values, typically as produced by METAL. Values are converted internally to -log10(P) for ranking signals.
N	Sample size.
radius	Maximum gap (bp) allowed between consecutive variants when defining a locus.
collapse.hla	Logical flag indicating whether variants within the HLA region should be collapsed.
build	Genome build used to define HLA boundaries ("hg19" or "hg38").

Details

Variants are merged into non-overlapping loci using valr::bed_merge(max_dist = radius). Variants connected through a chain of gaps less than or equal to radius belong to the same locus.

The most significant variant (largest -log10(P)) within each merged locus is selected as the lead signal.

When collapse.hla=TRUE, signals within the HLA region are collapsed using HLA boundaries:

• hg19/GRCh37: chr6:28477797-33448354

• hg38/GRCh38: chr6:28510020-33480577

Value

A tibble containing one lead variant per non-overlapping genomic locus.

Loci are defined by single-linkage clustering of variants separated by at most radius base pairs. Each locus is represented by its most significant variant (largest -log10(P)).

Output columns

• chrom: chromosome

• start, end: position of the lead variant

• rsid: variant identifier

• a1, a2: effect and other allele

• EAF: effect allele frequency

• b: effect size

• SE: standard error

• mlog10p: -log10(P) used for ranking

• n: sample size

• regionStart, regionEnd: boundaries of the merged locus

Notes

Each row corresponds to a non-overlapping locus defined using valr::bed_merge(max_dist = radius).

The lead variant is selected as the SNP with the highest mlog10p.

Examples

## Not run: 
threshold <- log10(5*10^-8)
varlist <- c(
  "Chromosome","Position","MarkerName","Allele1","Allele2",
  "Freq1","FreqSE","MinFreq","MaxFreq",
  "Effect","StdErr","log10P","Direction",
  "HetISq","HetChiSq","HetDf","logHetP","N"
)
tbl <- "~/Caprion/analysis/METAL_dr/ZPI_dr-1.tbl.gz"
sumstats <- read.table(tbl, col.names = varlist, check.names = FALSE)
res1 <- qtlFinder(sumstats)
res1
subset(res1,mlog10p>=-threshold)
d <- subset(sumstats, log10P <= threshold)
res2 <- qtlFinder(d)
res2
sentinels <- "~/Caprion/analysis/METAL_dr/sentinels/ZPI_dr.p.gz"
d2 <- read.table(sentinels, col.names = varlist, check.names = FALSE)
res3 <- qtlFinder(d2)
res3
# A tibble: 2 × 14
# Groups:   chrom, regionStart, regionEnd [2]
  chrom     start       end rsid   a1    a2      EAF      b     SE mlog10p     n
  <chr>     <dbl>     <dbl> <chr>  <chr> <chr> <dbl>  <dbl>  <dbl>   <dbl> <int>
1 chr13 113800622 113800622 13:11… t     c     0.373 -0.541 0.0306    69.0  2491
2 chr14  94750486  94750486 14:94… t     c     0.989  1.52  0.136     28.5  2491
# 3 more variables: regionStart <dbl>, regionEnd <dbl>, .overlap <int>

## End(Not run)

gap documentation built on Oct. 27, 2026, 1:07 a.m.