Table 1 NETGEM Generative process
From: NETGEM: Network Embedded Temporal GEnerative Model for gene expression data
Require: H {set of functional categories} |
|---|
Require: G = (V, E) {interaction network graph} |
Require: Λ ={λe}e∈E{prior on functional category mixtures} |
Require: Θ ={θ h }h∈H{prior on class transition probability} |
1: for all h in H do |
2: Choose Q
h
|θ
h
s.t. each row |
3: for all e in E do |
4: Choose |
5: for t = 1 to T do |
6: for all e in E do |
7: Choose |
8: Choose |
9: Choose |
10: Compute |
11: Choose |
where 
in (5)
in (1)- Table 1 shows the generative process for the gene expression data
for multiple strains s ∈ {1, ..., S} and observation time points t ∈ {1, ..., T} in the NETGEM model. The active functional category 
controls the change of interaction strength from 
to 
in edge e at time t. The interaction strengths affect the observed gene expression data, through the strain damping model and the conditional probability distribution in (1).
for multiple strains s 
controls the change of interaction strength from 
to 
in edge e at time t. The interaction strengths affect the observed gene expression data, through the strain damping model and the conditional probability distribution in (1).