Integration of Multiple Genomic Data Sources in a Bayesian Cox Model for Variable Selection and Prediction

Figure 10

Glioblastoma data: posterior selection probabilities for the models based on the uninformative (-axis) versus the informative (-axis) priors. The highlighted data points correspond to the selected variables with the largest posterior selection probabilities based on the two priors (blue diamonds identify variables selected only with the uninformative prior, red squares only with the informative prior, and purple triangles with both priors; see Figure 9).