Skip to content

Latest commit

 

History

History
40 lines (28 loc) · 1.27 KB

README.md

File metadata and controls

40 lines (28 loc) · 1.27 KB
Raw

NetMIM

Overview

NetMIM is a Bayesian framework integrating gene expression, DNA methylation and pathway information to identify significant biological features associated with disease outcomes. Our data includes gene expression, DNA methylation, a map from DNA methylation probes to genes, structure dependencies among genes and clinical covariates and response.

User manual

The required R packages:

  • Rcpp
  • RcppArmadillo
  • RcppGSL
  • scales
  • caret
  • MASS

After constructing the list data and list prior as shown in example.R, you can call NIBAMM function by

> source("R/NetMIM.R")
> result <- NIBAMM(data, prior, max_iters = 10000)

The full MCMC chain is stored in result, which can be used in the following analysis and plots. For example, we can find PPI of variables by

> n.burnin = 2000
> max_iters = 10000
> colMeans(result$gamma[n.burnin:max_iters, ])

If the response is patients' survival time, the log transformation with natural base is required firstly. Then the data can be constructed as shown in example.R. In this case, run MCMC by

> source("R/survivalNetMIM.R")
> result <- survivalNIBAMM(data, prior, max_iters = 10000)

If there is any bugs without being fixed timely, please contact author directly [email protected]