Skip to content
master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
R
 
 
 
 
man
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Bayesian deconvolution model for intracellular component exploration (BayICE)

BayICE: A hierarchical Bayesian deconvolution model with stochastic search variable selection.

To install this package, please copy and paste following codes into your R session:

  1. install.packages("devtools")
  2. library(devtools)
  3. install_github("AshTai/BayICE")
  • Required packages:
  1. require(MCMCpack)
  2. require(invgamma)
  3. require(quadprog)
  4. require(MASS)
  5. require(tibble)
  6. require(mvtnorm)

Example

library(BayICE)
data("NSCLC")
# Get the gene-specific expression for generating simulated data from real cancer data
n.base <- rowMeans(mdata.u[,grep("N",colnames(mdata.u))])
n.base <- n.base[n.base>1 & n.base < 1000]

sim.data <- MN_sim(n.base=n.base)

# For sequencing data, we recommend the log-transformation of raw data as the BayICE input.
# BayICE outputs the average of iterative data after burn-in and thinning. 
# If the complete MCMC data is required, please use BayICE_iter.
res <- BayICE(ref.set=log(sim.data$ref+1), mix.set=log(sim.data$mix+1),ref.id=sim.data$type,iter=2000,burnin = 0.6,
  thinning = 3)

Reference

An-Shun Tai, George C. Tseng, and Wen-Ping Hsieh* (2021). BayICE: A Bayesian hierarchical model for semi-reference-based deconvolution of bulk transcriptomic data. Annals of Applied Statistics. 15(1) 391 - 411. https://doi.org/10.1214/20-AOAS1376.

Contact information

An-Shun Tai (daansh13@gmail.com) https://anshuntai.weebly.com

About

BayICE: A hierarchical Bayesian deconvolution model with stochastic search variable selection

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages