Skip to main content

Table 1 Comparison of the fits of mathematical models A and B to the experimental data

From: Combining laboratory and mathematical models to infer mechanisms underlying kinetic changes in macrophage susceptibility to an RNA virus

Pig (Batch) Nr data points RMS model A RMS model B BIC model A BIC model B R 2 adj model A R 2 adj model B Bias model A Bias model B
1 (1) 25 0.088 0.026 −31.9 −58.4 0.87 0.99 0.071 0.005
2 (1) 25 0.068 0.024 −27.4 −55.8 0.91 0.98 0.003 0.055
3 (1) 25 0.070 0.036 −36.9 −46.1 0.92 0.97 −0.027 0.008
4 (2) 35 0.178 0.090 −29.4 −50.0 0.42 0.86 0.284 −0.035
5 (2) 35 0.135 0.048 −37.8 −69.1 0.43 0.92 0.276 −0.008
6 (3) 35 0.194 0.133 −26.7 −38.1 0.07 0.53 0.415 −0.033
7 (3) 35 0.196 0.111 −26.4 −43.8 0.11 0.73 0.689 −0.002
8 (3) 35 0.205 0.107 −25.1 −44.9 0.08 0.73 0.541 −0.034
  1. RMS (Root Mean Square difference), BIC (Bayesian Information Criterion), R 2 adj and total Bias were calculated according to Eqs. (1) to (4), respectively. Nr data points is the number of incubation periods for each experimental replicate (5 for batch 1 and 7 for batches 2 and 3) multiplied by five corresponding to the five fitting criteria ((i) the proportions of CD163 positive cells from the mock infected samples, and for the infected samples, the proportions of (ii) non-infected CD163 positive cells, (iii) non-infected CD163 negative cells, (iv) infected CD163 positive cells and (v) infected CD163 negative cells, respectively). For model B, equal switching rates for CD163 positive and CD163 negative cells (i.e. AS1 described below) were assumed as this produced the model of best fit (see section below)