qqhex {GGtools} | R Documentation |
obtain qqplot coordinates for the specific case of comparing a given distribution to that of multiple realizations from a permutation distribution, and bin these coordinates using hexbin, useful for very large samples
binnedQQ(dt, nxbins=20, ylim=c(0,76), xlim=c(0,30), end45=5, thrs=c(0,.001,.005,.01,.05), tempmar = c(6,4,4,5), ...) qqhex(sco, p1, p2, p3, fdr, nxbins = 20, thrs = c(0, 0.001, 0.005, 0.01, 0.05)) binqq(qqob, ylim = c(0, 76), xlim = c(0, 30), end45=5, ...)
dt |
a |
sco |
numeric vector of observed statistics |
p1 |
realization of null distribution for |
p2 |
realization of null distribution for |
p3 |
realization of null distribution for |
fdr |
vector of FDR associated with elements of |
nxbins |
number of bins to be used for samples from the null distribution |
thrs |
vector of thresholds in FDR to be used for ruling the plot |
qqob |
for binhex(), output of qqhex |
ylim |
vertical limits of rendering |
xlim |
horizontal limits of rendering |
end45 |
a segment is drawn from (0,0) to (end45,end45) to depict the line of identity |
tempmar |
numerical vector with 4 elements serving as a temporary setting of
the |
... |
not currently used |
for qqhex, a list with elements
hb |
output of |
thrs |
vector of input |
scothrs |
vector of observed statistics corresponding to FDRs in |
opar = par(no.readonly=TRUE) set.seed(123) x = c(rchisq(9000,1), rchisq(1000,12)) nn = lapply(1:3, function(x) rchisq(10000,1)) fd = pifdr(x, unlist(nn)) qqh = qqhex(x, nn[[1]], nn[[2]], nn[[3]], fd) par(mar=c(4,4,4,7)) binqq(qqh,xlim=c(0,10), ylim=c(0,20)) mtext(4, "FDR") par(opar)