[R] Error in t.test

[arun]

Hi Vivek,

res1 <- read.table("positve-res1.txt",header=TRUE)
 dim(res1)
#[1] 5292    4
#You could simplify the first loop as:
mean_Res2 <- as.vector(rowMeans(res1))
 identical(mean_Res1,mean_Res2)
#[1] TRUE


##Regarding the second loop:

pval_r2 <- as.vector(apply(res1,1,function(x) as.numeric(try(t.test(x)$p.value))))
length(pval_r2)
#[1] 5292
 sum(is.na(pval_r2))
#[1] 64


##From running your second loop

 length(pval_r1) ##stopped at the first instance where data are essentially constant
#[1] 63
 res1[64,]
#            p118_diff p132_1_diff p132_2_diff mean
#XLOC_000353         1           1           1    1


 head(which(is.na(pval_r2)))
#[1]  64  78  99 259 309 397



A.K.




On Thursday, January 9, 2014 9:58 AM, Vivek Das <vd4mmind at gmail.com> wrote:

I have some query. I would be glad if you can help me out. I am trying to calculate the p-value for a simple datafram, but am getting an error. The data in each column are relative differences.. The last column being the mean. I am giving the code snippet. I am not being able to figure out  why this error is coming up. 
My matrix looks like:


mean_Res1<-c()

for (i in 1:length(res1[,1]))
{
mean_Res1<-c(mean_Res1,mean(as.numeric(res1[i,])))

}


pval_r1<-NULL
for(i in 1:length(res1[,1])){
pval_r1<-c(pval_r1,t.test(as.numeric(res1[i,]))$p.value)
print(i)
}

Error in t.test.default(as.numeric(res1[i, ])) : 
  data are essentially constant


I am not being able to figure out why is the t-test failing? Here is my matrix in the attachment. Can you suggest me?



----------------------------------------------------------

Vivek Das