Table 1.

Summary of infection parameters.

Parameter	Description	Default value (unit)	Source
$\beta$	Cell infection rate	$1.13\times {10}^{-6}$ (${\text{h}}^{\text{-1}}$)	Baccam et al. (2006)
$\delta$	Cell death rate	$0.1$ (${\text{h}}^{\text{-1}}$)a	Baccam et al. (2006)
$p$	Virus growth rate	$9.9\times {10}^{-4}$ (${\text{h}}^{\text{-1}}$)b	Baccam et al. (2006)
$c$	Virus clearance rate	$0.125$ (${\text{h}}^{\text{-1}}$)	Baccam et al. (2006)
$\mu$	Per site mutation rate	$2.3\times {10}^{-6}$ ($\text{s/n/h}$)c	Sanjuán et al. (2010)
$\alpha$	Frequency of neut. mutations	$0.25$	Sanjuán, Moya, and Elena (2004)
$sH$	Harmonic mean of selection	$0.2$	Sanjuán (2012)
$G$	Genome size	$13588$ (bp)	NIAID
$\lambda$	Prob. of no del. mutation	$e^{-(1-\alpha )\mu G/s_{\text{H}}}$	see Equation (1)
$T_0$	Initial number of target cells	$4\times {10}^8$ (cells)	Baccam et al. (2006)
$V_{\text{inoc}}$	Inoculum size	10	McCaw et al. (2011) d

^aThe estimate from Baccam et al. (2006) is 0.167; however, this gives a cell generation time of 6 h ($1/\delta$), which is unrealistic for influenza A virus. Baccam et al. (2006) acknowledge this, and the model has an improved fit when a cell generation time of 12 h is used. Sanjuán (2012) calculates a cell generation time of 10 h ($\delta =0.1$) and notes that it is realistic for a range of eukaryotic viruses. Note that cell generation time is not necessarily equivalent to the viral generation time because many viruses release virions by budding prior to cell lysis.

^bThe estimate from Baccam et al. (2006) is $5\times {10}^{-4}$; however, it is more realistic to match the $R_0^{\text{wh}}$ value to the Baccam et al. (2006) estimate (see Equation 8), where $R_0^{\text{wh}}=11.1$.

^cThe reported value is $2.3\times {10}^{-5}$ substitutions per nucleotide site per cell generation (s/n/c) and we convert it to an hourly rate (s/n/h) using the cell generation time $1/\delta =10$ h.

^dReports the number of virions per transmission event inferred from ferret infection data as $4.3\pm 9.8$ virions.