Automated Biological Experimentation (ABEX) uses Gaussian Process models and Bayesian optimization to automatically design experiments for defined objective. Typically, Bayesian optimization is suitable for expensive-to-evaluate functions, a property that can almost always be attributed to the design of experiments for biological systems. Measurements of a biological system in a laboratory can be conducted for a range of conditions. We often do not fully understand the mechanisms through those conditions affect the measurable behaviour of the biological system. Therefore, an emulator of the input-output behaviour is constructed as a Gaussian Process. This is then used to evaluate an acquisition function that rewards prospective new evaluations of the system when there is some probability that a new optimum might be found.
The main case study tackled in this project is:
- Optimization of biological information processing. This case study attempts to select optimal concentrations of inducing chemical signals to a synthetic gene circuit that perceives those signals and produces fluorescent proteins. The circuits are extensions of the double receiver circuit from Grant et al. (Mol. Syst. Biol. 2016)
The installation process is described in great detail in our installation guide.
Visit the contribution guide.
### View the documentation
In the cloned ABEX repository:
cd docs
make html
And open _build/html/index.html (from ABEX/docs) using a web-browser.