Figure 1.

Modeling of causal relationship among variables. (a) We adopt the biologically meaningful definition of function, and treat conservation measures and biochemical signals as consequences. (b) The latent functional indicator Z is modeled as the parental variable and all the 22 annotations are treated as consequences. Also, we assume there is no direct causal relationship between any two annotations. Therefore the annotations are conditionally independent given Z. (c) Workflow of GenoCanyon functional prediction.