Differentiating between bites of 2 or 3 3 and 4 (Model 10 vs 14) significantly improves the fit of the model (additive hazard coefficient of 4 = 0

Differentiating between bites of 2 or 3 3 and 4 (Model 10 vs 14) significantly improves the fit of the model (additive hazard coefficient of 4 = 0.008, p-value = 0.036). purple = 1C10, green = 11C100, blue = 101C1000, red = 1000) had a distinct probability of contamination. For example model 6 is the full SR-2211 model with each sporozoite score using a different probability of contamination. The different grey boxes show consecutive residual-sporozoite scores which were grouped together. Model 1 shows the null model where the probability of contamination was the same for all those mosquitoes irrespective of the number of bites received. Models 2 and 3 shows the scenarios whereby the probability of contamination is determined solely by the number of bites or the number of infectious bites, respectively. The models are compared using Akaike information criterion (AIC) with the lower the value the more parsimonious the model (denoted in each column by *). Results are shown for humans with a pre-erythrocytic vaccine candidate (PEV). In mice the datasets with and without PEV are analysed separately and together (assuming resistance reduces the probability of contamination equally for bites with different residual-sporozoite scores). In all best fit models mosquitoes with a higher residual-sporozoite score had a higher contamination probability.(DOCX) ppat.1006108.s003.docx (23K) GUID:?0F1D29CC-EB5D-4D1D-8B8D-6D4E283F5AEE S4 Table: The semi-parametric components of the best-fit additive hazard models of humans and mice. Additive hazard regression coefficient show the additional hazard in days (see (16) for a full explanation). P-values indicate whether the coefficient is usually significantly different from zero. In humans overall p-values for the significance of residual sporozoite scores of 2,3 or 4 4 are not possible as their occurrence is usually correlated with one another (each volunteer received 5 bites of score 2 or above). However the significance of each of these scores can be seen by comparing Model 10 with Model 13 and 14 (see S3 Table for a list of models). Differentiating between bites of 2 or 3 3 and 4 (Model 10 vs 14) significantly improves the fit of the model (additive hazard coefficient of 4 = 0.008, p-value = 0.036). Equally differentiating between scores of 2 and 3 or 4 4 (Model 10 vs 13) showed SR-2211 sporozoite score was also significant here (additive hazard coefficient of 2 = -0.016, p-value = 0.010).(DOCX) ppat.1006108.s004.docx (18K) GUID:?866967A7-A858-400A-8D93-42AD0903F9A4 S1 Fig: A diagram illustrating two possible hypotheses for how the efficacy of a vaccine might be associated with the number of residual-salivary gland sporozoites. Panel A repeats Fig 1E from the main text showing how the probability of infection in mice changes with the number of residual-sporozoites (with an arithmetic scale on the x-axis instead of a logarithmic). The number of residual-sporozoites was categorised so the mid-point of each bin is plotted (with the exception of the highest 1000 which takes the value 1200). Colours denote the category of residual-sporozoites, be it 0 (orange), 1C10 (purple), 11C100 (green), 101C1000 (blue) or 1000 (red). Circular points show na?ve mice whilst squares show those given the anti-CSP antibody. The non-linear shape of the relationship indicates that CSF2 the probability of infection is a negative density-dependent process. Panel B uses a made up graph to illustrate how SR-2211 this non-linear relationship can cause the efficacy to change according to the number of residual-sporozoites. The solid black line denotes a hypothetical relationship between the probability of infection and the number of residual sporozoites (similar SR-2211 to that observed in Panel A) whilst the other lines give different hypotheses for the action of a pre-erythrocytic vaccine. The pink dashed line assumes that the vaccine reduces the per-sporozoite transmission probability by 50% (here referred to as a leaky dose response) whilst the brown dotted-dashed line assumes that the vaccine reduces the number of sporozoites surviving in the host by a constant amount (the same across all hosts) so no mosquito with 400 residual-sporozoites is capable of transmitting the infection (here referred to as a threshold dose response). Panel C shows how these two hypotheses influence the relationship between vaccine efficacy and the number of residual-sporozoites. Further work is required to differentiate between these hypothesis though the human data presented here is consistent with the threshold dose response as only bites with 1000 residual-sporozoites appeared to contribute to transmission.(DOCX) ppat.1006108.s005.docx (89K) GUID:?89747080-2B94-4528-A7C3-CFAF4C8BDA48 S1 Dataset: Raw mouse data used to fit the probability of infection and time to patency models. (XLSX) ppat.1006108.s006.xlsx (80K) GUID:?22E52F30-C7B4-4AA4-8AE9-F3617ED0CE65 Data Availability StatementAll relevant mouse data are within the paper and its Supporting Information files. Abstract Over a century since Ronald Ross discovered that malaria is caused by.