Univariate Cox analysis for a gene list

Asked May 18 '22 at 04:43

Active May 18 '22 at 06:09

Viewed 168 times

I'm trying to plot this: enter image description here

using this code form GitHub:

PP<- data.frame()
for (i in names(tt)) {
   dd <- tt[[i]]
   for(j in colnames(dd)[4:ncol(dd)]){
      scox <- summary( coxph(Surv(OS.time,OS) ~ get(j),dd))
      p <- compareC(dd$OS.time, dd$OS, dd$IRLS,dd[,j])$pval
      uni <- rbind(uni, data.frame(ID=i,A=j,
                            HR=scox$conf.int[,1],
                            HR.95L=scox$conf.int[,3],
                            HR.95R=scox$conf.int[,4],
                            pvalue=scox$coefficients[,5],
                            cindex=scox$concordance[1]%>%as.numeric(),
                            cse=scox$concordance[2]%>%as.numeric(),
                            CP=p))}}

how this loop works? my input is a gene expression vector (TCGA) and i wanted to get a list of significant prognostic features (genes).

thanks

edited May 18 '22 at 06:09

IRTFM

258,963
21
364
487

asked May 18 '22 at 04:43

pyrosequence

Welcome to SO! Please share your data for reproducibility with `dput(dd)`. – YH Jang May 18 '22 at 05:53
Thanks dear Jang, I'm not getting what you exactly mean. may you state more clear. – pyrosequence May 18 '22 at 06:08
I mean I want you to share your data using `dput(your data)` in `R` to help your problem :) https://stackoverflow.com/a/49995752/11926169 – YH Jang May 18 '22 at 06:13
With no loop indices on the LHS of assignments you will overwrite the values of all calulations except the last ones. – IRTFM May 18 '22 at 06:16
oh okay, my data.frame is like this:600 samples in columns and 500 genes in rows – pyrosequence May 18 '22 at 07:05

Univariate Cox analysis for a gene list

0 Answers0