raPlot {caroline}

Generate a Ratio Average [RAy] Plot.

Description

A plot which turns two vectors of count data into log scaled fold change ratio and average abundance. The plot derives from a Bland-Altman plot and is also very similar to an MA plot. The RA plot is unique, however, in it's creative inclusion of the vector-unique 'arms' which are artificially introduced into the plot by adding a <1 epsilon factor before the log function is applied. The name RAy comes from the fact that the aforementioned 'uniques' arms addition makes it strongly resemble a geometric ray. Many of the parameters to the function play off of this convenient anatomical analogy.

Usage

raPlot(a, b=NULL, uniques=5, normalize=FALSE,  
        nr=0, alpha = 0.01, jitter=FALSE, jit.wgts=NULL,
        rex=1, flat=TRUE, tail=.5, arms=.5, spine=1, border=NULL, plot=TRUE,  ...)

Arguments

a

a vector of counts for a. can also be a matrix with two columns 1 for a and 2 for b.

b

a vector of counts for b.

uniques

a boolean specifying whether or not to plot the library-unique genes (those with zero counts in one or the other library).

normalize

A boolean specifying whether or not to normalize the counts into proportions.

nr

a numeric value indicating the asymptotic normalization ratio line.

alpha

a statisical significance value.

jitter

whether or not or how much to jitter the a and b counts into surrounding, non-overlapping space.

jit.wgts

a weight vector used to spread the counts of a and b into surrounding, non-overlapping space.

rex

a numeric value specifying the radial expansion of the plot points.

flat

a boolean for the radial expansion of points as a function of both R and A axes.

tail

a numeric or boolean value indicating the line thickness of the two trailing curved significance lines of the RAy.

arms

a numeric or boolean value indicating the line thickness of the two leading straight separator lines of the RAy.

spine

a numeric or boolean value indicating the line thickness of the normalization line (whose y position is specified by mm).

border

a vector of strings used to color the borders of the points.

plot

whether or not to do the actual plot.

...

other parameters passed to plot.

Value

a RAy plot

See Also

limma::plotMA, edgeR::maPlot

Examples

a <- rnbinom(n=10000, mu=5, size=2)
b <- rnbinom(n=10000, mu=5, size=2)

## the alternative 
plot(a,b)
## the raPlot version
raPlot(a, b) 

## highlight the condition unique points in the same way as  edgeR's "maPlot"
RA <- raPlot(a, b, pch='')
cond.unique <- apply(cbind(a,b), 1, function(d) any(d==0))
points(RA$A,RA$R, col=c('black','orange')[cond.unique+1])

## try playing with jittering over plotted points
raPlot(a, b, jitter=.3)