Entering edit mode
Erick Okuto
▴
10
@erick-okuto-5787
Last seen 10.5 years ago
Dear all,
I have a data with rows as my pixels and columns as my samples. These
pixels are expected to be distributed differently hence cannot be
treated
as if they are similar. I need to normalize them using the available
procedures in affy package or any other that may do it better.
Unfortunately, the data is in raster format and affy package require
matrix
of intensities. Also, raster files are in large numbers therefore
conversion to other formats tend to increase the computational cost in
terms of time.
Hypothetical example:
Require(raster)
r <- raster(ncols=36, nrows=18)
n <- 1000
x <- runif(n) * 360 - 180
y <- runif(n) * 180 - 90
xy <- cbind(x, y)
# get the (last) indices
r0 <- rasterize(xy, r)
How do we normalize r0 so that rows may be similar using the existing
methods in affy package or any other recommended procedure. Please
help.
This is just a typical example which would then enable me to write a
function
that allow simultaneous runs for multiple raster images. Many thanks
indeed
for your assistance.
Cheers,
Erick Okuto, Kenya.
[[alternative HTML version deleted]]