Our overall inference framework is limited by several assumptions at each of the modeling levels. Regarding the molecular docking, while the output is constrained by the common caveats of the approach itself such as the ability of the scoring functions to forecast accurate binding energies, findings from both experimental and previous docking studies are aligned with our findings. As for the in vitro estimates, we assumed that the viral dynamics in the experimental environment approximated those in vivo; this important caveat can likely only be tested further in clinical studies. However, we used ranged efficacy estimates aiming to produce more flexible scenarios. Further, Zandi et al. [14] reported that TDF did not inhibit SARS-CoV-2 was not inhibited by TDF when using concentrations under 20 microM. Interestingly, in this study Vero cells were not preincubated with TDF in contrast with the procedure from Clososki et al. where cells were preincubated with the drug 24h before virus inoculation, which might explain the discrepancy and aligns with our model predictions.