My thesis is titled “Graph Transformation, Atom Tracing, and Isotope Labelling” and takes place in the intersection of Computer Science and Chemistry, called Cheminformatics.
My thesis describes how we can first model molecules and chemical reactions as graphs and graph transformations in a framework called MØD. My task is then to trace atoms through the chemical network to see where they end up.
The classical example: When you eat glucose sometimes (unfortunately) it is turned into fat molecules. This process is largely understood and we know that there are two ways this can happen: The ED pathway and the EMP pathway. But in some species, ED is active and in others the EMP is active. The question is: For a given species, which one?
One method that the scientists can use is to buy a special version of glucose where one of the carbons is not the usual carbon-12, but rather a carbon-13 (a different isotope). Imagine this as being colored differently from the rest of the carbons. These special molecules can then be passed through the digestive system of some species and we can see where the carbon-13 atoms end up, in particular in what position in the fat molecules they end up. This is the basis for isotope labelling experiments.
The problem is that these isotope labelled compounds, such as the special glucose, are orders of magnitude more expensive than the regular and the scientists might have to buy many different versions where the carbon-13 is in different positions of the glucose to actually learn something interesting.
My task is to take first steps towards automating this, so a computer simulation can be used instead, in order to save time and money.
In my thesis I take a group theoretical approach which has not been done before. It is an interesting combination of chemistry, graph theory and abstract algebra.
You can read my thesis here: PDF
The slides used for my defence: PDF