Table 3.

Estimated relative entry rates in the DD recruitment model (θ₀ and θ₁) strongly depend on the value of assumed exit rate from the cluster μ.

	μ=0.06/h	μ=0.1/h	μ=0.3/h	μ=1/h	μ=3/h
Estimated θ0	1.45 (1.10–1.88)	0.92 (0.70–1.23)	0.41 (0.30–0.51)	0.24 (0.17–0.31)	0.20 (0.14–0.28)
Estimated θ1	6.31 (4.37–7.71)	4.02 (2.77–4.91)	1.77 (1.29–2.09)	1.04 (0.84–1.17)	0.89 (0.77–0.98)

We fixed the value of the rate of exit of T cells from the cluster μ to different values [indicated in the top row) and fitted the DD recruitment model (with λ_k = λ₀+λ₁k and μ_k = kμ, see Equations (1), (2)] to experimental data on clustering of PyTCR T cells (n = 130) when transferred alone (cluster formation was observed 6 h after T cell transfer, see Figure 5A). The quality of model fits to data as judged by the negative log-likelihood were nearly identical between different fits ($\mathcal{L}~\mathit{\text{194}}$); the fit of the model with μ = 3/h is shown in our previous publication [see Figure S1A and Table S1 in Cockburn et al. (29)]. To compare parameter estimates we show relative entry rates θ₀ = λ₀/μ and θ₁ = λ₁/μ. In parentheses we shown 95% confidence intervals for parameter estimates. Interestingly, the ratio θ₁/θ₀ was relatively constant for different fits (53, and manuscript in preparation).