Figure 4.

Evolutionary dynamics of advantageous mutants. (A) Fixation probability as a function of the infection probability, B. The blue line with diamonds shows the fixation probability in the absence of multiple infection, provided by formula (1). The green line shows the prediction of formula (2). The closed black circles show the fixation probabilities observed in the agent based simulation when one cell infected with one mutant virus is introduced into the system at equilibrium. The black open circles show the fixation probabilities observed in the computer simulation when one mutant virus is randomly placed into any of the available cells in system at equilibrium. The inset re‐plots the observed fixation probability shown in closed black circles, and the red crosses depict the prediction given by formula (3) when the composite fitness value r’ is calculated according to formula (4), as described in the text. Parameters were: B₁ = 0.025, B₂ = rB₁, A = 0.02, L = 1, D = 0.01, N = 900, r = 1.0005. The number of simulation results for black closed circles was: 4,866,352; 5,371,603; 3,577,510; 4,091,305; 2,648,860; 1,691,486; and 1,272,341. For black open circles: 3,935,759; 4,490,230; 3,313,452; 4,309,990; 3,470,896; 4,837,538; and 5286726. (B and C) Same simulations, but with larger mutant advantages, r = 1.001 for (B) and r = 1.01 for (C). For (B), the number of simulation runs for the closed black circles are; 7,553,368; 6,920,249; 6,011,459; 5,067,733; 3,155,326; 1,939,997; and 1,507,107 . For black open circles: 17,417,910; 19,829,787; 14,557,581; 18,888,450; 15,088,002; 11,334,612; and 19,407,499. For (C), closed black circles: 7,519,349; 6,572,315; 5,445,510; 4,494,334; 2,831,424; 1,767,559; and 1,362,813. For (C), black open circles; 854,795; 773,698; 702,491; 681,172; 547,988; 407,116; and 333,571. Trends described in the text are statistically significant, according to the Z test for two population proportions. (D) Average time until 90% of the infected cell population contain the advantageous mutant for the first time (black closed circles), based on the agent‐based model with mutations and back‐mutations, as a function of the infection probability. Standard errors are plotted, but are hard to see. The numbers of simulation runs are: 19,839; 41,663; 82,828; 222,263; 316,597; 638,422; and 488,754. The blue line depicts the result of equivalent simulations in the absence of multiple infection. Again, standard errors are too small to see, and the number of simulation runs are: 130,825; 226,054; 282,790; 481,975; 494,045; 1,080,864; and 998,265. Parameters were: B₁ = 0.025, B₂ = rB₁, A = 0.02, L = 1, D = 0.01, μ = 3 × 10⁻⁵, N = 900, r = 1.01. The trends described in the text are statistically significant, according to the two‐sample t‐test.