Date of Award

Spring 2014

Document Type


Degree Name

Doctor of Philosophy (PhD)


Mathematics, Statistics and Computer Science

First Advisor

Corliss, George

Second Advisor

Merrill, Stephen

Third Advisor

Spiller, Elaine


The leading malaria vector control strategies (i.e., long-lasting insecticidal nets and indoor residual spraying) can reduce indoor transmission, but these tools alone are insufficient to eliminate it. Strategies that target adult mosquitoes when they feed on humans or animals outdoors or target mosquito immature stages are also needed to achieve malaria elimination. Improved data systems for integrating

diverse experimental observations and research groups, as well as process-explicit mathematical models for evaluating them are both essential to achieving these goals.

We have developed a generic schema and data repositories for the studies of malaria vectors that encompass a wide variety of different experimental designs that rapidly generate large data volumes. We extended a malaria transmission model to examine the relationship between transmission, control, and the proportion of blood

meals a vector population obtains from humans: Assuming the lower limit for this indicator of human feeding preference enabled derivation of simplified models for zoophagic vectors. We present differential equation models to describe the biological processes that mediate novel strategies to control malaria vectors by autodissemination

of pyripoxyfen (PPF) as it is transferred from treated stations to the gravid

mosquitoes and then to the aquatic habitats where it inhibits mosquito emergence.

Data from most of the mosquito studies we reviewed conformed to our

generic schema with four tables recording the experimental design, sorting of collections, details of samples, and additional observations. Our corresponding online repository includes 20 experiments, 8 projects, and 15 users at two institutes, resulting in 10 peer-reviewed publications. For zoophagic vectors, the results from

model can be used to forecast the likely immediate and delayed impacts of an intervention using only three field-measurable parameters. For the autodissemination of PPF, sensitivity analysis indicates success of the strategy is plausible because the ≥ 80% coverage of aquatic habitats with PPF appears achievable with modest, biologically plausible values of field-measurable input parameters.

Therefore, we have applied two of the computational sciences aspects (i.e., research data preparation using computer systems and scenario analysis with mathematical models) to address obstacles to the control and elimination of malaria.