Skip to main content


Fig. 1 | Malaria Journal

Fig. 1

From: A novel model fitted to multiple life stages of malaria for assessing efficacy of transmission-blocking interventions

Fig. 1

A graphical outline of the multi-generational transmission experiment (a) and its mathematical representation (b). Ten days prior to treatment, 5 female TO mice (6–8 weeks old) were injected with 107–108 Plasmodium berghei clone ANKA 2.34. One day prior to the trial, mosquitoes were starved. Infected mice were either treated with the transmission-blocking intervention, atovaquone ATV (0.5 µg kg−1 in 100 µl dimethyl sulfoxide DMSO), or given a negative control (DSMO alone). After 2 h, mice were anaesthetized and 500 naïve An. stephensi (line sd 500) mosquitoes were allowed to feed on all 5 mice simultaneously and at random so any mosquito could feed on any mouse. This cohort of 500 mosquitoes was sub-sampled (n = up to 50 mosquitoes) and dissected 10 days after feeding to measure the prevalence and intensity of oocysts (O’). The mosquitoes were at their most infectious 21 days after feeding on the infected mice [8]. At this point, a group of 5 naïve mice (N) were anaesthetized and fed to a specified number of mosquitoes (m = 1–5 mosquito bites per mouse). Immediately after successful feeding (determined by an engorged abdomen), the sporozoites (S) remaining within the mosquito post-feeding were then scored on a binned scale (representing either 0, 1–10,11–100, 101–1000, or 1000 + sporozoites per mosquito). Over 10 days, the bite-exposed mice were sampled for parasitaemia and gametocytaemia. These mice were then treated with the control or ATV and a new cohort of 500 naïve mosquitoes were allowed to feed on any mouse simultaneously at the start of the second transmission cycle (i = 2 cycles). The transmission cycles were repeated four times. This experiment was reported previously in [12]. b The probabilistic Bayesian model mirrored the experimental set up. The initial parasite density generated by injecting mice N 0, the oocyst intensity O’ and the parasite density in mice transmitted by mosquito bites N, are modelled assuming zero-inflated negative binomial distributions. The sporozoite count S data for the biting mosquitoes were censored which means that the data are modelled as a multinomial distribution. These distributions are defined by the mean (µ, o, s for parasite density in mice, oocyst counts and sporozoite counts in mosquitoes) and dispersion (φ, τ, σ for parasite density in mice, oocyst counts and sporozoite counts in mosquitoes) and zero-inflation (π P , π V ) parameters. Parameters from the previous life stage are used to inform the next (the respective parameter informing the subsequent life-stage is indicated by the arrows). The biting effect m is modelled when sporozoites in mosquitoes propagate parasite infections in mice for each transmission cycle i and treatment arm t. All care and handling of animals strictly followed the Guidelines for Animal Care and Use prepared by Imperial College London, and was performed under the UK Home Office Licence 70/7185

Back to article page