Bacteriophages benefit from generalized transduction
Fig 3
Identification of relevant parameters for phage survival.
A) Random Forest Analysis (RFA) is based on 5000 randomized combinations of parameter values and 20 repeated simulations for each combination. Parameters with a higher % in increased minimum square error have greater impact on the measured outcome: the fraction of simulations (for the 20 replicate simulations with a similar parameter combination) where phage are found alive. The parameters varied in the RFA are shown in S1 Text, section 3. B) Detailed effects of 3 variables in the survivability of initially antibiotic sensitive bacteria. The outcomes of simulations are shown in function of the probability of generalized transduction (y-axis) and genome size of the antibiotic resistant bacteria on which the phage was propagated (first panel), phage burst size (second panel) and number of phage particles sampled from the first part of the experiment (third panel) used for the second part of the experiment (see Fig 2). Each bin corresponds to the median of 20 replicate simulations for each parameter combination.