sThresh {NanoStringNCTools} | R Documentation |
Convenience functions for matrix thresholding, centering, and scaling based upon margin statistics.
# Loop over features fThresh(x, STATS) fCenter(x, STATS) fScale(x, STATS) ## Round results to integers fIntThresh(x, STATS) fIntCenter(x, STATS) fIntScale(x, STATS) ## Comparisons fAbove(x, STATS) fBelow(x, STATS) fAtLeast(x, STATS) fAtMost(x, STATS) # Loop over samples sThresh(x, STATS) sCenter(x, STATS) sScale(x, STATS) # Round results to integers sIntThresh(x, STATS) sIntCenter(x, STATS) sIntScale(x, STATS) ## Comparisons sAbove(x, STATS) sBelow(x, STATS) sAtLeast(x, STATS) sAtMost(x, STATS)
x |
a numeric array. |
STATS |
the summary statistic for thresholding, centering, or scaling. |
These functions are convenience wrappers for the following code:
fThresh
:sweep(x, 1L, STATS, FUN = "pmax")
fCenter
:sweep(x, 1L, STATS, FUN = "-")
fScale
:sweep(x, 1L, STATS, FUN = "/")
fIntThresh
:round(sweep(x, 1L, STATS, FUN = "pmax"))
fIntCenter
:round(sweep(x, 1L, STATS, FUN = "-"))
fIntScale
:round(sweep(x, 1L, STATS, FUN = "/"))
fAbove
:sweep(x, 1L, STATS, FUN = ">")
fBelow
:sweep(x, 1L, STATS, FUN = "<")
fAtLeast
:sweep(x, 1L, STATS, FUN = ">=")
fAtMost
:sweep(x, 1L, STATS, FUN = "<=")
sThresh
:sweep(x, 2L, STATS, FUN = "pmax")
sCenter
:sweep(x, 2L, STATS, FUN = "-")
sScale
:sweep(x, 2L, STATS, FUN = "/")
sIntThresh
:round(sweep(x, 2L, STATS, FUN = "pmax"))
sIntCenter
:round(sweep(x, 2L, STATS, FUN = "-"))
sIntScale
:round(sweep(x, 2L, STATS, FUN = "/"))
sAbove
:sweep(x, 2L, STATS, FUN = ">")
sBelow
:sweep(x, 2L, STATS, FUN = "<")
sAtLeast
:sweep(x, 2L, STATS, FUN = ">=")
sAtMost
:sweep(x, 2L, STATS, FUN = "<=")
An array with the same shape as x
that has been modified by
thresholding, centering, or scaling.
Patrick Aboyoun
# Find reasonable column minimums thresh <- apply(stack.x, 2L, quantile, 0.05) # Threshold column values identical(sThresh(stack.x, thresh), sweep(stack.x, 2L, thresh, FUN = "pmax")) # Substract column values identical(sCenter(stack.x, thresh), sweep(stack.x, 2L, thresh)) # Scale to common mean identical(sScale(stack.x, colMeans(stack.x) / mean(colMeans(stack.x))), sweep(stack.x, 2L, colMeans(stack.x) / mean(colMeans(stack.x)), FUN = "/")) # Scale to common mean, rounded to the nearest integer sIntScale(stack.x, colMeans(stack.x) / mean(colMeans(stack.x)))