Skip to main content


Figure 3 | Genetics Selection Evolution

Figure 3

From: A unifying theory for genetic epidemiological analysis of binary disease data

Figure 3

Effect of including different sources of host variation on the prediction of individual disease status. ROC curves calculated with data from simulation of the SIR model with variation in susceptibility and infectivity; the classification criterion used was the probability function equation (14) with Λjt including different sources of variation; curves in green = ‘Genetic epidemiology’ - Λjt includes all sources of variation and was estimated based on equation (15) (AUC = 0.964); orange = ‘Epidemiology’ - Λjt was estimated assuming no (genetic) variation in susceptibility and infectivity, as in equation (17) (AUC = 0.895); purple = ‘Genetics’ - Λjt was estimated assuming (genetic) variation in susceptibility and infectivity, but equal non-dynamic exposure, i.e. D k t = D ¯ for each individual k (AUC = 0.710); black = random classification (AUC = 0.5); grey = perfect classification (AUC = 1).

Back to article page