Package 'fastqq'

Internal function to prune quantiles of non-important values for visualization.

Description

This function is not exposed, since we want to hard-code the parameters for simplicity of usage.

Usage

drop_dense(x, y, N_hard = 10000)

Arguments

x	A numeric vector of sample/theoretical points.
y	A numeric vector of theoretical/sample points.
N_hard	Desired upper bound on the number of points to plot.

Value

data.frame with o and e pruned as columns.

---

Creates a Q-Q plot

Description

Creates a quantile-quantile plot from p-values from an association study, e.g. a genome wide association study (GWAS). We compare the data quantile with a theoretical quantile from a uniform distribution. This code is mostly adapted from the qqman package, but improved for speed. A graph with a hundred million points should only take a few seconds to generate.

Usage

qq(pvector, ...)

Arguments

pvector	A numeric vector of p-values.
...	Other arguments passed to plot()

Value

No return value, called for plotting side effects.

Examples

qq(stats::runif(1e6))

---

Creates a Q-Q plot for comparing with normal quantiles

Description

Faster alternative to stats::qqnorm(). For more than 1e5 points we remove excess points, that would not be visible in the plot, since the points are so close. Otherwise this should work exactly the same, and the code is mostly adapted from stats::qqnorm(). This code produces more lightweight plots for excessive amounts of data.

Usage

qqnorm(
  y,
  ylim,
  main = "Normal Q-Q Plot",
  xlab = "Theoretical Quantiles",
  ylab = "Sample Quantiles",
  plot.it = TRUE,
  datax = FALSE,
  ...
)

Arguments

y	sample, to compare to normal quantiles.
ylim	graphical limits.
main	Plot title.
xlab	X label.
ylab	Y label.
plot.it	Should the plot be created.
datax	logical. Should data values be on x-axis?
...	Other arguments passed to plot()

Value

data.frame with sorted sample and normal quantiles, NA values are excluded.

Examples

qqnorm(stats::rnorm(1e6))

---

Creates a Q-Q plot

Description

Faster alternative to stats::qqplot(). For more than 1e5 points we remove excess points, that would not be visible in the plot, since the points are so close.

Usage

qqplot(
  x,
  y,
  plot.it = TRUE,
  xlab = deparse1(substitute(x)),
  ylab = deparse1(substitute(y)),
  ...
)

Arguments

x	First sample for qqplot.
y	Second sample for qqplot.
plot.it	Should the plot be created.
xlab	x label for plot.
ylab	y label for plot.
...	Other arguments passed to plot()

Value

list with sorted samples, interpolated to be same size.

Examples

qqplot(stats::runif(1e6),stats::runif(1e6))

---