Table 2. Model comparisons using Bayes factors (likelihood ratios of two competing models) and biological interpretations.
Genotype | BF ± 1 SE | Interpretation of BF |
---|---|---|
A. Immune response required? Comparing hybrid model with Mideo et al. (2008b) model: |
||
AS + DK | (5.5 ± 3.0) × 1052 | Immune responses required to explain data |
B. Age-structure required? Comparing hybrid model with Miller et al. (2010) model: |
||
AS + DK | .44 ± .08 | Age structure not required to explain data |
C. Removal of extra death rate for multiply parasitized RBCs: |
||
AS + DK | (1.1 ± .2) × 1016 | Increased death rate required to explain data |
D. Removal of immune-mediated merozoite clear- ance from model: |
||
AS | (1.1 ± 1.1) × 1016 | Decisive evidence of clearance of AS merozoites |
DK | 1.0 ± .1 | No evidence of clearance of DK merozoites |
E. Removal of immune-mediated parasitized RBC clearance from model: |
||
AS | 3.6 ± .4 | Some evidence of parasitized RBC clearance |
DK | 2.8 ± .7 | Some evidence of parasitized RBC clearance |
F. Removal of immune-mediated unparasitized RBC clearance from model: |
||
AS | 4.9 ± .5 | Some evidence of unparasitized RBC clearance |
DK | .40 ± .05 | No evidence of unparasitized RBC clearance |
Note: Values are Bayes factors (BF) ± 1 standard error of the BF. A, Original model from Mideo et al. (2008b) versus hybrid model incorporating age structure and immunity. B, C, Comparisons of hybrid model with reduced versions. D–F, Effects of removing key immune components from hybrid model. RBC = red blood cell.