I recently used proggene for survival analysis, which majorly uses NCBI-GEO data. Is anyone familiar about the usage or implementation of the risk summary that the web tool provides at the end of analysis? I understand that proggene itself works by using R based analysis, but I am trying to use the risk summary from here to manually plot R graphs, so as to adjust the no of days in my analysis. Any suggestions will be appreciated.
Using median gene expression value as bifurcating point, samples are divided into High and Low gene expression groups. Using survival data and continuous expression variable, survival analysis is done by fitting cox proportional hazards model using function “coxph” of library survival. Hazard ratio (HR) as 'exp(coef)’ and log rank p value are retrieved from the fitted model. To create prognostic plot, High and Low expression categorical variable is used along with survival data. Plots are created using function 'survfit’ of the same R library.
As stated, they use the 'survival' package in R and there are a number of good tutorials with this package. See my previous reply to similar question here
