Fig. 6

Expected relatedness under superinfection. Bootstrap distributions for the expected relatedness under superinfection were generated by randomly sampling with replacement 31 monogenomic pairs. For each set of 31 monogenomic pairs, we calculated the average relatedness and repeated this process 40,000 times to generate bootstrapped distributions of the mean relatedness between monogenomic infection pairs. Superinfection was simulated with either a simple random sampling scheme (blue), in which all sample pairs were equally likely, or a weighted sampling scheme (green), which uses the barcode frequencies of the corresponding monogenomic samples to weigh each sample pair. Bootstrap resampled distributions of expected relatedness in polygenomic infections are shown in orange. p values for both sampling schemes were ≤2.5 × 10⁻⁵