Fig 1.
a. Schematic of bacteriophage lytic cycle, characterised by seven life history parameters. b. Assembly curve, showing the number of viable virions within an infected cell as a function of time: initially zero until the eclipse period elapses, then increasing linearly. c to e. Assembly curves, each showing the effect of a faster-than-average lysis cycle (purple) versus the average (red), assuming the cycle is faster because the eclipse period is shorter (c), the post-eclipse is shorter (d), or both are proportionally shorter (e).
Table 1.
Viral life history parameters, their units and definitions.
Fig 2.
Schematic description of the agent-based model’s implementation.
Healthy cells and progressing growth clocks are pictured yellow. Infected cells, viable virions and progressing lysis clocks are red. Virions that have injected their DNA, or that have decayed are pictured grey, as well as bacterial growth clocks which have halted due to infection.
Table 2.
Parameters used in serial passage simulations.
Table 3.
Parameters used in plaque expansion simulations.
Table 4.
Bacteria parameters used in all simulations.
Table 5.
Bacteriophage parameters used in numerical analysis and simulations.
Fig 3.
Dependence of bacteriophage growth rate on lysis time under optimal conditions.
a. Bacteriophage effective growth rate k as a function of mean lysis time, with either 0 minutes or 2.5 minutes standard deviation in lysis time. Inset: example lysis time distributions with mean 17 minutes, std 0 (solid line) or 2.5 minutes (dashed line). b. Optimal lysis time under ‘linear with eclipse’ model as a function of lysis time std. c. Effective growth rate achieved at optimal lysis time, under ‘linear with eclipse’ model, as a function of lysis time std. d. to f. Effective growth rate as a function of both mean lysis time and lysis time std for ‘normal’, ‘’linear‘, and ‘linear with eclipse’ models respectively.
Fig 4.
Results of serial passage competition experiment, competing a wild type against a mutant phage with different mean lysis times.
i. Difference in numerically predicted growth rate, using Eq 3. ii. Mutant fixation probability in stochastic simulation. iii. Fitness difference in stochastic simulation. Note that in the first and third rows, the visual scale has been set to saturate at .
Fig 5.
Results of serial passage competition experiment, in which a wild type phage with lysis time mean 17 min, std 2.5 min competes against a mutant with varied lysis time mean (y-axis) and std (x-axis).
i. Difference in numerically predicted growth rate, using Eq 3. ii. Mutant fixation probability in stochastic simulation. iii. Fitness difference in stochastic simulation. Note that in the first and third rows, the visual scale has been set to saturate at . Axes are scaled such that each pixel step in the y direction corresponds to a 1 minute change in mean lysis time, while each pixel step in the x direction corresponds to an equivalent fractional change in lysis time std versus the wild type value: 2.5/17 minutes. The four-point star on all panels correspond to the point in parameter space at which both phages have exactly the same parameters.
Fig 6.
a. Plaque expansion experiment schematic. b. Results of plaque expansion competition experiment, in which a wild type phage with lysis time mean 17 min, std 2.5 min competes against a mutant with varied lysis time mean (y-axis) and std (x-axis). i. Difference in numerically predicted growth rate, using Eq 3. ii. Mutant fixation probability in stochastic simulation. iii. Fitness difference in stochastic simulation. iv. Infection front velocity. Note that in the first and third rows, the visual scale has been set to saturate at . Axes are scaled such that each pixel step in the y direction corresponds to a 1 minute change in mean lysis time, while each pixel step in the x direction corresponds to an equivalent fractional change in lysis time std versus the wild type value: 2.5/17 minutes. The four-point star on all panels corresponds to the point in parameter space at which both phages have exactly the same parameters. The white polygons in rows ii and iv outline the parameter space in which the wild type reaches fixation.