Table 1.

Parameter estimates for both constant size and exponential population growth models

	Constant			Exponential
Zebra no.	θ	Ne	TMRCA	θ0	β	Ne (1)	TMRCA	CFU
1	1.05	286.16	88.14	1.9(1.4,2.63)	0.69(0.2,1.18)	215.94(120.18,528.12)	1.47(0.86,4.35)	1.71(0.95,4.17)
2	1.08	294.08	89.41	1.92(1.39,2.7)	0.74(0.23,1.23)	212.13(116.52,514.11)	1.39(0.83,3.94)	1.68(0.92,4.06)
3	0.79	215.5	67.18	1.88(1.38,2.62)	0.57(0.1,1.08)	232.67(118.23,634.15)	1.74(0.94,7.33)	1.84(0.93,5.01)
5	0.52	142.3	44.13	1.88(1.38,2.61)	0.36(0.1,0.89)	295.38(126.63,636.11)	2.61(1.13,7.22)	2.34(1.00,5.03)
7	1.09	297.81	90.85	1.92(1.38,2.72)	0.76(0.25,1.26)	209.15(114.66,508.35)	1.36(0.81,3.75)	1.65(0.91,4.02)
8	1.03	280.55	82.05	2.02(1.27,3.33)	0.96(0.29,1.6)	193.25(88.07,621.59)	1.11(0.64,3.37)	1.53(0.70,4.91)
9	0.6	163.24	49.7	1.91(1.31,2.88)	0.58(0.1,1.18)	235.22(101.56,708.49)	1.74(0.86,7.23)	1.86(0.80,5.60)
13	1.1	301.08	91.46	1.92(1.38,2.75)	0.77(0.25,1.28)	208.01(112.63,509.58)	1.33(0.8,3.69)	1.64(0.89,4.03)
14	0.52	142.3	44.13	1.88(1.38,2.61)	0.36(0.1,0.89)	295.38(126.63,636.11)	2.61(1.13,7.22)	2.34(1.00,5.03)
17	0.53	145.22	45.06	1.89(1.38,2.66)	0.45(0.1,1)	260.27(116.15,650.11)	2.15(1.01,7.21)	2.06(0.92,5.14)
19	0.28	77.24	24.26	2.42(1.67,3.64)	1.62(0.63,2.47)	118.09(52.91,375.73)	0.73(0.47,1.81)	0.93(0.42,2.97)

$\theta$ is the average number of mutations that separates two genes under the coalescent process. It is defined as twice the effective population size N_e times the mutation rate µ. This number remains the same under the constant effective population size model. Under the exponential population growth model, the zebra’s B. anthracis population value of $\theta$ at the moment of death is ${\theta }_0$, and the effective population size changes (from present to past) according to the exponential function $N_e\left(t\right)=N_e\left(0\right){\text{e}}^{-\beta t}$, where $\beta$ is the exponential rate parameter and $N_e\left(0\right)={\theta }_0/2\mu$. Accordingly, N_e (1) represents the effective population size of B. anthracis in each zebra at the moment of infection using the experiment’s estimated mutation rate (see full model and statistical analyses description in Methods). Confidence intervals are calculated only for the exponential population growth model, since it was the best fit to the data. TMRCA is expressed in days, assuming a mutation rate of 0.002. The founding size of the population has been converted to CFU from the effective population size from the exponential model.