Contents

visualize$genome$overview

One-call genome overview of a PRS library

Compute the default genome signals for a model library and stack their tracks in one call: the applied shortcut for the compute-then-stack workflow, readable at any library size.

Usage

visualize.genome.overview(
  models,
  gwas = NULL,
  model.associations = NULL,
  outcome = NULL,
  reference = NULL,
  attribution = FALSE,
  traits = NULL,
  top.n = NULL,
  region = NULL,
  bands = NULL,
  min.models = 2,
  max.labels = 40,
  heights = NULL
)

Arguments

Argument	Description
`models`	A PGSLibrary or a single PGS. Any other class, or an empty set, aborts.
`gwas`	A single-variant PolyGeniusAssociation from [associate$singleVariant()](/reference/associate-single-variant/), or NULL (default). Adds the Manhattan and convergence tracks and anchors concordance to the outcome; NULL omits both tracks. Recognized by a chr/position column on $results (the usual aliases accepted); a per-model regression passed here aborts by name rather than being silently ignored.
`model.associations`	A per-model PRS regression PolyGeniusAssociation from [associate$regression()](/reference/associate/), or NULL (default, sum over the whole library). Gates the convergence track to the FDR-significant models. Recognized by its predictor column; a single-variant result passed here aborts by name.
`outcome`	Character scalar, or NULL (default). Outcome selected in every association input; required when one of them holds more than one.
`reference`	Character scalar naming a model in models, or NULL (default). Used only when gwas is absent, where it anchors the concordance sign to that index model; NULL anchors to the canonical allele, leaving only the amount of agreement meaningful. List the choices with names(models$models).
`attribution`	Logical scalar, default FALSE. Add the per-trait attribution lanes, which are off by default because they are unreadable at large library sizes. Has no effect without gwas.
`traits`	Character vector of trait names and a positive integer count, or NULL (default) for each. Narrow the attribution lanes, as for [visualize$genome$attribution](/reference/visualize-genome-attribution/); ignored when the attribution track is not drawn.
`top.n`	Character vector of trait names and a positive integer count, or NULL (default) for each. Narrow the attribution lanes, as for [visualize$genome$attribution](/reference/visualize-genome-attribution/); ignored when the attribution track is not drawn.
`region`	A "chr:start-end" string, or NULL (default) for the whole genome. Passed through to the stack.
`bands`	Intervals to shade through every panel, or NULL (default). Passed through to the stack, which documents the accepted forms.
`min.models`	Numeric scalar, default 2. Minimum number of carrying models for a variant to enter the concordance signal.
`max.labels`	Numeric scalar, default 40. Row-label cap for the attribution lanes.
`heights`	Numeric vector, or NULL (default, each track's own height). Relative panel heights, one per resulting track. The track count depends on which signals could be built, so set heights on a manual [stack](/reference/visualize-genome-stack/) instead when it matters.

Value

A patchwork composite, as visualize$genome$stack returns, with one panel per track that could be built. Apply a theme to every panel with &, not +.

Which tracks appear

Top to bottom: a coverage rail, showing where the library places variants; a single-variant Manhattan and a convergence track, both only when gwas is given; a concordance track, always; and the per-trait attribution lanes last, only when attribution = TRUE and gwas is given. Convergence is gated to the FDR-significant models when model.associations is also given. A track whose signal cannot be produced is dropped with a warning rather than failing the figure; if every track fails, all the recorded errors are raised together.

Examples

visualize$genome$overview(models, gwas = gwas, outcome = "dementia")
visualize$genome$overview(models, gwas = gwas, attribution = TRUE, top.n = 10)