Home

Last updated: 2016-05-10

Code version: 20c8f5a7774c1b3712f5a383545be34a95b086fa

Analyzed Simulated Data

Note that the below analyses are done with CONDITIONING on the covariance matrix \(\Sigma\). The reason for this is that we can include a shrinkage term for indirect effects and still have Bayes Factor be analytic with summary data.

• Simulated data WITHOUT shrinking indirect effects
• Simulated data WITH shrinking indirect effects
• Real Simulated data WITHOUT shrinking indirect effects
• Real Simulated data WITH shrinking indirect effects 1
• Real Simulated data WITH shrinking indirect effects 2
• Check Shrinkage With Oracle Plugin

Analyzed Real Data

For now, these data were analyzed unconditional on \(\Sigma\), using an improper Wishart prior for \(\Sigma\).

• Adipose Tissue, 0.80 correlation cutoff
• Muscle Skeletal, 0.80 correlation cutoff
• Whole Blood, 0.80 correlation cutoff

Check Model with Real Data

We would expect gene networks to be similar in similar tissues. All networks had a maximum of 10 genes for a correlation cutoff of 0.60.

• Skin, 0.60 correlation cutoff