The MICRF (mutation integration of Conditional Random Fields) model based on R and Matlab.
Matlab code is mainly based on UGM toolbox to train the best parameters for each biological networks. R code is mainly based on CRF R package to infer and decode the risk labels for each genes.
Simulated data are running for TADA firstly.
coexp.R: building the co-expression networks from BrainSpan microarray data.
Idea: see report Qiang_3_24.pptx. The integrated data can find more risk genes than the sum of the single data set.
7_6 and 3_17 report slide for old version model details.