Skip to contents

Screen a GDS for candidate sex-linked markers

Source: R/sex_markers.R
sexy_markers.Rd

sexy_markers() compares marker presence, heterozygosity, and, when available, normalized read depth between known females and males. Dominant SilicoDArT markers stored with MARKER_TYPE = SILICODART contribute presence/absence evidence but are not interpreted as diploid heterozygosity. It is a read-only screen: active GDS sample and variant selections are respected and restored, and no filter is applied or written to the GDS.

Usage

sexy_markers(
  data,
  strata = NULL,
  sex.column = "STRATA",
  presence.threshold = 0.9,
  absence.threshold = 0.1,
  min.heterozygosity.difference = 0.2,
  coverage.ratio.threshold = 1.5,
  coverage.threshold = 1,
  min.samples.per.sex = 5L,
  fdr.threshold = 0.05,
  require.significance = FALSE,
  chunk.size = 2000L,
  save.plots = TRUE,
  plot.formats = c("png", "pdf"),
  folder.name = NULL,
  verbose = TRUE,
  ...
)

Arguments

data

A GDS filepath or an open SeqVarGDSClass object.

strata

A data frame or tab-delimited file containing INDIVIDUALS and sex.column. If NULL, individual metadata are read from the GDS.

sex.column

Metadata column containing sex. F, female, M, and male are recognized without regard to case; other values are unknown.

presence.threshold

Minimum presence in the expected sex for Y/W.

absence.threshold

Maximum presence in the other sex for Y/W.

min.heterozygosity.difference

Minimum absolute female-minus-male heterozygosity difference for X/Z candidates.

coverage.ratio.threshold

Minimum larger-to-smaller normalized depth ratio for X/Z coverage candidates.

coverage.threshold

Minimum read depth considered present. Without depth data, a non-missing genotype is considered present.

min.samples.per.sex

Minimum number of known females and males.

fdr.threshold

Maximum FDR when significance is required.

require.significance

Require effect size and method-specific FDR.

chunk.size

Number of variants read from the GDS at a time.

save.plots

Save diagnostic plots in the result folder.

plot.formats

One or both of "png" and "pdf".

folder.name

Optional result-folder stem. Default: sexy_markers.

verbose

Display progress and summary messages.

...

Common arguments: path.folder and internal.

Value

A sexy_markers object containing marker statistics, candidates, an assignment-ready Y/W panel, sample summaries, a metadata audit, plots, and the result-folder path.

Details

Y-like markers are present mainly in males, W-like markers mainly in females, X-like markers have greater heterozygosity or normalized depth in females, and Z-like markers show the reverse. These labels are candidates, not proof of chromosomal location. Sex-specific dropout, population structure, plate or library effects, paralogy, and mapping bias can produce similar patterns.

Statistical tests

Presence and heterozygosity use two-sample score tests for proportions. Normalized depth uses Welch tests on log2(1 + depth). P-values are adjusted separately by method with Benjamini-Hochberg FDR. Candidate selection always uses explicit effect-size thresholds; require.significance = TRUE also requires the method-specific FDR to pass fdr.threshold.

Read-depth normalization

When depth is available, each sample is divided by its mean positive depth across active markers. This reduces sample-wide sequencing-depth differences but cannot remove marker-by-batch interactions or confounding between sex and plate, lane, library, population, or sampling group.

See also

Run vignette("detecting_sex_linked_markers", package = "radr") for the biological background, guidance on discovery data, SilicoDArT and VCF workflows, interpretation, alternative methods, and validation.

Examples

if (FALSE) { # \dontrun{
result <- radr::sexy_markers(
  data = "study.gds",
  strata = "sample_metadata.tsv",
  sex.column = "SEX"
)
result$candidates
} # }