pQTL Mediation analysis

Description Usage Arguments Details Examples

Plot LOD statistics calculated by [mediation_scan()] against index using ggplot2.

ggplot_mediation_scan(x, col = "firebrick4", cex = 1,
  xlab = index_name, ylab = "Conditioned LOD", col_target = "blue",
  gap = 25)

## S3 method for class 'mediation_scan'
autoplot(x, ...)

## S3 method for class 'mediation_scan'
plot(x, ...)

mediation_scan(target, mediator, driver, annotation, covar = NULL,
  intcovar = NULL, kinship = NULL, method = c("double-lod-diff",
  "ignore", "lod-diff"), fitFunction = fitDefault, facet_name = "chr",
  index_name = "pos", verbose = TRUE)

`x`	mediation object
`col`	color of points (default "firebrick4")
`cex`	character expansion (default 2)
`xlab, ylab`	X and Y axis label (default 'index_name' and "Conditioned LOD")
`col_target`	color for target LOD line
`gap`	gap between facets (default '25')
`target`	A numeric vector with gene/protein expression
`mediator`	A matrix, each column is one gene/protein's expression
`driver`	A matrix, haplotype probabilities at QTL we try to mediate
`annotation`	A data frame with mediators' annotation with columns 'facet_name' and 'index_name'
`covar`	A matrix with additive covariates
`intcovar`	A matrix of covariate interacting with driver
`kinship`	kinship object
`method`	A method to handle missing cases
`fitFunction`	function to fit models
`facet_name`	name of facet column (default 'chr')
`index_name`	name of index column (default 'pos')
`verbose`	If TRUE display information about the progress

For a given QTL haplotype probabilities 'driver“ and target 'target', the function sequentially tries to add each column of 'mediator' matrix as a covariate and calculates LOD statistic. The low LOD value indicates 'driver' and 'target' are conditionally independent given 'mediator', i.e. 'mediator' is a mediator of causal relationship from 'driver' to 'target'.

data(Tmem68)
target <- Tmem68$target
m <- match("Tmem68", Tmem68$annotation$symbol)

med_scan <- mediation_scan(target = target,
                      mediator = Tmem68$mediator,
                      driver = Tmem68$qtl.geno,
                      annotation = Tmem68$annotation,
                      covar = Tmem68$covar,
                      method = "double-lod-diff")
ggplot2::autoplot(med_scan)
ggplot2::autoplot(subset(med_scan, "4")) +
  ggplot2::geom_vline(xintercept = Tmem68$annotation[m,"pos"], linetype = "dashed")