bim.fdr {bim} | R Documentation |
Bayesian false discover rate (FDR) applied to QTL mapping. Use plot.bim.fdr to determine FDR for HPD level used in plot.bim.effects.
bim.fdr(x, cross, nqtl = 1, pattern = NULL, exact = FALSE, chr, ..., levels, df = 3, qtl ) plot.bim.fdr(x, cross, ..., fdr, critical.value, hpd = NULL )
x |
object of class bim |
cross |
corresponding object of class cross |
nqtl |
subset on number of QTL |
pattern |
subset on chromosome pattern of QTL |
exact |
subset on exact pattern or number of QTL if true |
chr |
subset of chromosomes to plot (numerical indices or chromosome names) |
levels |
probability levels for HPD regions (default every 1 percent) |
df |
degrees of freedom for smoothing spline estimator of prior on size |
qtl |
QTL estimate information from bim.qtl |
fdr |
object returned by call to bim.fdr |
critical.value |
pFDR guidelines (ignored if hpd provided; default every .05 to .25) |
hpd |
HPD guidelines |
... |
graphical parameters can be given as arguments to
plot |
FDR follows ideas of Storey (http://www.stat.berkeley.edu/~storey) adapted to Bayesian setting.
bim.fdr
returns an enhanced version of the
bim.qtl
object, with additional elements:
levels |
HPD levels |
size |
relative size of HPD region across levels |
fdr |
false discover rate across levels |
hyp |
estiamtes for probabilities of hypotheses H0: no QTL at locus, M0: no QTL anywhere, M1: QTL somewhere |
prob |
cumulative prior probability |
spline |
spline estimate of cumulative prior probability |
hyp |
estiamtes for probabilities of hypotheses H0: no QTL at locus, M0: no QTL anywhere, M1: QTL somewhere |
fdr |
FDR for corresponding HPD levels (chosen by hpd or
critical[value settings) |
Brian S. Yandell, yandell@stat.wisc.edu
http://www.stat.wisc.edu/~yandell/qtl/software/bmqtl
data( vern ) data( verngeo.bim ) plot.bim.fdr( verngeo.bim, vern, pattern=c(1,1) )