Why did we model? What did we hope to achieve?
We generated mathematical models of our DNA-based system, our protein-based system, and the effects associated with the implementation of our diagnostic device on a human population. The aim of our models was to better understand our systems and improve them to become more feasible in practice.
We also created a piece of software, which highlights the versatility of our device – by changing some of the DNA sequences (particularly the cleavage sites and output), our device can be used to detect various diseases. Our software aims to make it easier for other groups to use our diagnostic for other diseases.
What techniques did we use for modelling?
We employed numerous modelling techniques and equations to most accurately assess our systems. These include mass action and Michaelis-Menten kinetics; sensitivity analyses; parameter scans; stochastic modelling; and SIR and vector epidemiological modeling. Please follow the links below to see more detailed explanations of our models and to try out our software!
How is our model integrated with the rest of our project?
Our modelling was closely interlinked with the rest of our project: it greatly informed and was informed by our design, wet lab work, and human practices. Some examples include:
- The epidemiological modeling confirmed that focusing our project and device on the diagnosis of congenital Chagas disease can have a profound impact on the levels of the disease in a population. It also informed our applied design and contributed to our decision to make a Chagas public health poster.
- The modelling of our DNA-based system led us to confirm the need for a strong ribosome binding site (RBS) to minimize false negatives
- We communicated with experts in blood clotting, Dr Scott Diamond and Professor Mike Laffan. They provided incredibly important information that enabled us to model the output of our device and modify our design accordingly.
- DNA-based system models with a hirudin output indicated to us that we would need an amplification step to limit false negatives: instead of producing hirudin in response to cruzipain, we decided to produce TEV protease, which can cleave multiple molecules of already-present, inactive hirudin.
- In our efforts to create a complex and accurate epidemiological model, we consulted experts in the field, including Professor Mike Bonsall.
- When designing our protein-based system parts, we decided to tag our fusion proteins with sfGFP so future experimental data could better inform our modelling, for example with respect to efficiency of protein transport to the outer membrane. Unfortunately, however, we encountered difficulties with the cloning of some of our OMV-targeted proteins and were unable to include experimental parameters in our models.
For more detail, please look at the specific modelling pages!