Unifying Viral Genetics and Human Transportation Data to Predict the Global Transmission Dynamics of Human Influenza H3N2

Lemey, Philippe; Rambaut, Andrew; Bedford, Trevor; Faria, Nuno; Bielejec, Filip; Baele, Guy; Russell, Colin A.; Smith, Derek J.; Pybus, Oliver G.; Brockmann, Dirk; Suchard, Marc A.

cause they reflect the frequency at which the predictor is included in the model and therefore represent the support for the predictor. As opposed the analysis reported in main manuscript (Fig. 2) , which specifies a prior probability of 0.019 on each predictor's inclusion, we here specify a prior probability of 0.5 on the inclusion of each predictor. Indicator expectations corresponding to Bayes factor support values of 3 and 20 are shown as a thin and thick vertical line respectively in these bar plots. The contribution of each predictor, when included in the model (bDd~1), where b is the coefficient or effect size, is represented by the mean and credible intervals of the GLM coefficients on a log scale. NA 1 : no conditional effect size available because the effect was never included in the model. We tested different population size and density measures, different incidence-based measures and different seasonal measures (Text S1), but only list the estimates for a representative predictor for the sake of clarity. NA 2 : no indicator expectation or conditional effect size available because the predictor was not available for this discretization of the sequence data. A comparison with the analysis reported in main manuscript (Fig. 2) indicates that our results are robust to the prior specification for the inclusion probabilities; only the scale of the Bayes factor support shifts to lower values because of the higher prior odds (1:1 as opposed to 0.019:0.981) in this case. (PDF) Figure S4 Net Markov jump counts for the 14 air communities. For each air community, we summarize the average net Markov jumps (jumps to -jumps from) and their 95% credible intervals. The estimates are ordered from the lowest (top; jumps to ,jumps from) to highest (bottom; jumps to .jumps from) net jumps. The data points are colored according to the air communities represented in Fig. 1 in the main text. (PDF) Figure S5 Correlation among observed H1N1 peaks and simulated peaks based on the BSSVS estimates. The Spearman rank correlation (r) and mean absolute error (MAE; in days) for all locations except for Mexico is shown at the top left. The data points are colored according to the air communities represented in Fig. 1 in the main text. (PDF)