although other posts had the same question, i couldnt use the solution. I am trying to generate a matrix for estimate of correlation for only the significant values. It is supposed to be simple but i am getting an error "Error in mat[i, j] <- result["estimate"] : incorrect number of subscripts on matrix" Here is my GENE input:
Name Sample1 Sample2 Sample3 Sample4 Sample5 Sample6 Sample7 Sample8 Sample9 Sample10 Lrriq3 8.185794 5.691456 5.693373333 6.973468667 8.868912 5.915211333 6.718336667 6.212762667 6.424637333 13.01974667 Dnase2b 0 0.1749128 0 0.1685122 0.1784736 0.122940127 0.007396118 0 0 0.09347276 Lphn2 1.080010133 10.01754067 14.10849333 11.77894 1.2552028 1.702124667 11.52506 15.21622 0.093035673 0.019666988 Rpf1 7.439926667 8.863518 10.28811467 11.86218 13.45304667 13.44146667 20.04024 16.94706667 23.76358 17.00742667 Uox 7.458356667 10.01754067 14.10849333 11.77894 19.75814 12.14829333 14.58846667 11.52506 15.21622 14.57954 Ctbs 0.400568 0.134638993 3.450422667 0.164317553 0 0 0.3395462 0.079734033 0.2700658 0 Spata1 2.066878 2.079750667 1.7238 2.240882667 1.461403333 2.093744 1.67564 1.2552028 1.702124667 1.427768 Ptprh 1.080010133 0.09089988 0.621011133 0.3004404 0.228991467 0.063827739 0.188904267 0.093035673 0.256751333 0.424108067
My LNC input:
Name Sample1 Sample2 Sample3 Sample4 Sample5 Sample6 Sample7 Sample8 Sample9 Sample10 XX1 3.956263333 2.443864667 1.413482 1.486519333 2.20473 3.015326 1.1033612 0.977534 0.789298267 1.469496 XX2 2.759029333 2.371987333 3.434 4.004905333 5.198814667 2.889342 3.463316 4.039935333 5.038084667 5.113266667 XX3 4.214811333 3.470377333 8.075684667 5.115368 7.084812667 4.767865333 6.272181333 6.202424667 5.480058667 4.613682 XX4 3.256852667 2.944397333 2.047966 1.696964667 2.099414667 1.780854667 0.3989612 0.23245 0.257986867 1.676498 XX5 661.7403333 647.749 834.8288 670.8856 728.8326667 710.5224667 357.7705333 387.9334 404.3672667 694.4849333 XX6 7.458356667 10.01754067 14.10849333 11.77894 11.77894 19.75814 11.77894 1.2552028 1.702124667 11.52506 XX7 7.458356667 10.01754067 14.10849333 11.77894 19.75814 14.58846667 11.52506 13.45304667 13.44146667 0.23245
Script is designed to do correlation between each row from each file (note that the samples 1 to 10 are arranged in same order) and output an excel sheet with the p value, estimate, and test, as well as a matrix of the estimates for only those with p <0.05. All the script works except for one step.
Script is:
genes <- read.delim(file="SampleGene.txt", header=TRUE, row.names=1)
lnc <- read.delim(file="Samplelncs.txt", header=TRUE, row.names=1)
x = rownames(genes[1:nrow(genes),])
y = rownames(lnc[1:nrow(lnc),])
d<-NULL #creates an empty dataframe
mat<-matrix(0,nrow(genes),nrow(lnc)) #creates a matrix with all values as 0
rownames(mat) <- rownames(genes) #assigns rownames to the matrix based on row names of the gene file
colnames(mat) <- rownames(lnc) #assigns colnames to the matrix based on the colnames of the lnc file
for (i in x){
for (j in y) {
result=cor.test(as.numeric(genes[i,]), as.numeric(lnc[j,]), method='pearson')#calculates the correlation and assigns it to result
d<-rbind(d, data.frame(i, j, result[c("estimate","p.value","statistic","method")], stringsAsFactors=FALSE)) #rbind allows writing output of loop to an empty dataframe. Works perfectly.
if (result["p.value"]<0.05){ #attempts to add the estimate to the matrix only of p.value <0.05
mat[i,j] <- result["estimate"] #This is causing the error
#print(result["estimate"]) #if I just print without adding to matrix, i dont get errors
}
}
}
write.table(file="Pearson.xls", as.data.frame(d), sep="\t")
As I indicated, if i remove the if statement from the loop OR if i just print out result["estimate"], I dont get errors. Otherwise, I get errors al the time.
I am a beginner in R and programming.Hence, if there are other suggestions to optimize the script above, please let me know.
d<-NULL
does NOT create an empty dataframe. – IRTFM