you could try :
trt1 <- do.call("rbind", lapply(listexp, function(x) {
zz <- aggregate(t(x), list(rep(1:3, each=2)), mean)[,-1]
zz <- as.matrix(zz)
sweep(zz[1:2,], 2, zz[3,], "-")[1,]}))
trt2 <- do.call("rbind", lapply(listexp, function(x) {
zz <- aggregate(t(x), list(rep(1:3, each=2)), mean)[,-1]
zz <- as.matrix(zz)
sweep(zz[1:2,], 2, zz[3,], "-")[2,]}))
expeffect <- list(trt1=trt1, trt2=trt2)
expeffect <- lapply(expeffect, function(x) {names(x) <- names(listexp) ; x})
szhan at uoguelph.ca a ??crit :
Hello, Everyone,
I am sorry that my message got truncated due to wrong format.
I hope it works now:
Hello, R Users,
I have a list (say listexp) of 10,000 elements, each of which consists of a
matrix (5X6). It likes:
$"a"
trt1rep1 trt1rep2 trt2rep1 trt2rep2 ctlrep1 ctlrep2
[1,] 50 54 98 89 40 45
[2,] 60 65 76 79 34 43
[3,] 86 83 34 45 38 34
[4,] 67 78 88 98 45 41
[5,] 55 59 77 88 56 66
$"b"
trt1rep1 trt1rep2 trt2rep1 trt2rep2 ctlrep1 ctlrep2
[1,] 55 54 88 86 42 40
[2,] 66 65 86 99 44 48
[3,] 80 86 44 55 38 44
[4,] 77 78 98 92 35 41
[5,] 50 53 87 88 46 56
.
I want to perform some operations on the list and then got a new list(say
expeffect) like this:
1. first average two replicates for each row in the matrix shown above (like
treatment 1 using (trt1rep1+trt1rep2)/2) and get one value for each
treatment/control
2. then subtract average value of the control from the each treatment and
one value for the treatment effect for each row in the matrix shown above
3. make a new list (say trteffect) with 2 elements (trt1 and trt2), each of
which consists of a matrix (10,000X5)with a row name same with old list's
element name, a column name corresponding to the row name of old list's
like:
$"trt1"
[,1] [,2] [,3] [,4] [,5]
a 9.5 24 48.5 29.5 -4
b 13.5 19.5 42 39.5 0.5
.
$"trt2"
[,1] [,2] [,3] [,4] [,5]
a 51 39 3.5 50 21.5
b 46 46.5 8.5 57 30.5
.
Could you please help me to make this new list (expeffect)?
Thank you in advance!
Joshua