Table 1.
Linear mixed effect models selected for 12 immune parameters.
| Immune trait | Best model selected | Age function | Variable | Parameter estimate ± SE | t-value | p | R2m | R2c | Predicted value at two years | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ♂ TF | ♀ TF | ♂ CH | ♀ CH | |||||||||
| INNATE TRAITS | ||||||||||||
| Neutrophil count | I(age^2)*sex + pop | quadratic | Intercept | 6.00 ± 0.25 | 23.97 | *** | 0.09 | 0.51 | 6.25 ± 0.29 | 6.52 ± 0.15 | 5.34 ± 0.34 | 5.14 ± 0.18 |
| I(age^2) | 0.01 ± 0.003 | 3.97 | *** | |||||||||
| Sex (M) | 0.57 ± 0.29 | 1.96 | — | |||||||||
| Pop (CH) | −1.16 ± 0.22 | −5.15 | *** | |||||||||
| I(age^2): sex (M) | −0.02 ± 0.01 | −3.69 | *** | |||||||||
| Monocyte count | Constant | — | Intercept | 0.29 ± 0.09 | 3.12 | — | 0.00 | 0.46 | 0.30 ± 0.10 | 0.31 ± 0.10 | 0.29 ± 0.07 | 0.30 ± 0.10 |
| Basophil count | Constant | — | Intercept | 0.07 ± 0.02 | 4.33 | — | 0.00 | 0.28 | 0.07 ± 0.02 | 0.07 ± 0.02 | 0.07 ± 0.02 | 0.08 ± 0.01 |
| Eosinophil count | Sex | — | Intercept | 0.12 ± 0.01 | 10.03 | *** | 0.02 | 0.15 | 0.07 ± 0.01 | 0.12 ± 0.02 | 0.10 ± 0.02 | 0.11 ± 0.01 |
| Sex (M) | −0.03 ± 0.01 | −2.90 | ** | |||||||||
| Hemagglutination | Age * pop | threshold (8 years) | Intercept | 2.43 ± 0.65 | 3.76 | *** | 0.01 | 0.44 | 4.03 ± 0.33 | 4.03 ± 0.25 | 4.26 ± 0.35 | 3.98 ± 0.36 |
| Age | 0.19 ± 0.07 | 2.85 | ** | |||||||||
| Pop (CH) | 2.67 ± 0.88 | 3.04 | ** | |||||||||
| Age: Pop | −0.32 ± 0.10 | −3.01 | ** | |||||||||
| Hemolysis | Age * pop | threshold (10 years) | Intercept | −1.52 ± 1.31 | −1.17 | — | 0.01 | 0.62 | 2.24 ± 0.41 | 2.35 ± 0.37 | 2.10 ± 0.40 | 1.97 ± 0.38 |
| Age | 0.38 ± 0.12 | 3.07 | ** | |||||||||
| Pop (CH) | 5.11 ± 2.06 | 2.48 | * | |||||||||
| Age: Pop | −0.53 ± 0.20 | −2.57 | * | |||||||||
| INFLAMMATORY TRAITS | ||||||||||||
| Alpha1-globulin | Age + I(age^2) + pop + sex | quadratic | Intercept | 3.35 ± 0.19 | 17.45 | *** | 0.04 | 0.5 | 3.35 ± 0.20 | 3.11 ± 0.19 | 3.24 ± 0.16 | 2.94 ± 0.15 |
| Age | −0.09 ± 0.03 | −2.54 | * | |||||||||
| I(age^2) | 0.01 ± 0.003 | 3.29 | ** | |||||||||
| Pop (CH) | −0.13 ± 0.05 | −2.76 | ** | |||||||||
| Sex (M) | 0.18 ± 0.05 | 3.79 | *** | |||||||||
| Alpha2-globulin | Constant | — | Intercept | 5.79 ± 0.30 | 19.62 | − | 0.00 | 0.26 | 5.63 ± 0.30 | 5.84 ± 0.34 | 5.66 ± 0.33 | 5.88 ± 0.29 |
| Beta-globulin | I(age^2) + sex + pop | quadratic | Intercept | 5.84 ± 0.38 | 15.22 | *** | 0.12 | 0.43 | 6.20 ± 0.54 | 5.73 ± 0.36 | 7.12 ± 0.51 | 6.41 ± 0.41 |
| I(age^2) | 0.02 ± 0.002 | 8.71 | *** | |||||||||
| Sex (M) | 0.63 ± 0.16 | 3.81 | *** | |||||||||
| Pop (CH) | 0.64 ± 0.16 | 3.87 | *** | |||||||||
| Haptoglobin | Age * pop + sex | threshold (9 years) | Intercept | −4.24 ± 0.74 | −5.71 | *** | 0.07 | 0.15 | 0.34 ± 0.10 | 0.15 ± 0.06 | 0.69 ± 0.14 | 0.30 ± 0.11 |
| Age | 0.48 ± 0.08 | 6.04 | *** | |||||||||
| Pop (CH) | 3.18 ± 1.20 | 2.66 | ** | |||||||||
| Sex (M) | 0.25 ± 0.08 | 3.05 | ** | |||||||||
| Age: Pop | −0.33 ± 0.13 | −2.50 | * | |||||||||
| ADAPTIVE TRAITS | ||||||||||||
| Lymphocyte count | Age + I(age^2) + pop | quadratic | Intercept | 2.81 ± 0.17 | 16.49 | *** | 0.14 | 0.37 | 2.25 ± 0.14 | 2.42 ± 0.13 | 1.90 ± 0.17 | 1.89 ± 0.15 |
| Age | −0.15 ± 0.05 | −2.76 | ** | |||||||||
| I(age^2) | 0.01 ± 0.004 | 2.12 | * | |||||||||
| Pop (CH) | −0.65 ± 0.08 | −8.27 | *** | |||||||||
| Gamma-globulin | Age + pop | threshold (4 years) | Intercept | 12.32 ± 1.14 | 10.5 | *** | 0.21 | 0.65 | 13.86 ± 0.90 | 13.54 ± 0.93 | 18.55 ± 1.66 | 18.43 ± 1.57 |
| Age | 0.52 ± 0.08 | 6.79 | *** | |||||||||
| Pop (CH) | 4.40 ± 0.39 | 11.43 | *** | |||||||||
The effect of different age functions (factor, linear, threshold, quadratic), of sex (F: Female, M: Male), of population (TF: Trois-Fontaines, CH: Chizé), with all two and three-way interactions between them, were tested. All models included individual identity, the year of capture and the cohort of individuals as random effects. When the age threshold model was selected, we have indicated the age at which the parameter begins to vary, and the “Parameter estimate” of the age function is the slope of the variation with age after the threshold age. Statistical significance is represented by *for p < 0.05, **for p < 0.01 and ***for p < 0.001. R2m and R2c are the marginal and conditional variance of the model, respectively. Values are presented ± Standard Error.