Table 1.
Use of OD sums to globally evaluate the role of DH on antibody production.
| Tested effect |
P | B1 | B2 | B3 | B4 | Null model | Alternative model |
|---|---|---|---|---|---|---|---|
| G | 0.00056 | 0.349 | 0.0472 | 0.0433 | 0.187 | Y=A+E+ε | Y=A+E+G+ε |
| G*E | 2.00E-08 | 3.26E-08 | 7.71E-06 | 0.00132 | 0.00733 | Y=G+A+E+G*A+A*E+ε | Y=G+A+E+G*A+A*E+G*E+ε |
| G*A | 0.0199 | 0.269 | 0.0395 | 0.00618 | 0.00472 | Y=G+A+E+A*E+G*E+ε | Y=G+A+E+A*E+G*E+G*A+ε |
| G*A*E | 1.78E-15 | 3.21E-07 | 2.58E-05 | 0.00334 | 0.00014 | Y=G+A+E+A*E+G*E+G*A+ε | Y=G+A+E+A*E+G*E+G*A+G*A*E+ε |
OD (optical density) values in ELISAs are continuous and correlate with antibody concentration. To evaluate the global role of DH sequence on antibody production, the OD values at all dilutions for each sample were summed. Mixed effect models were used for comparisons. Variables used in the comparison included the effect of genotype (G), antibody type (IgM or IgG) (A), and epitope (E) after the prime immunization (P) and after each of the four booster immunizations (B1–B4). As per common practice in statistical analyses of this type, these studies considered both a null model and an alternative model, which contains the additional variable to be tested. E.g., when testing for genotype, the null model lacked this variable and the alternative model included it In addition to the fixed effects provided in the models, both included mouse as a random effect. The residual, ε, is the difference between the actual and predicted ODsums for each model.