Bayesian networks of data from a cohort study:
I am analyzing data from a cohort study to clarify the causes of asthma.
Simulation of the anti-cancer drug ABT-737:
ABT-737 is a new anti-cancer which readily induces apoptosis (cell death) in cancer cells but not in healthy ones. I am modelling the apoptosis-management signalling networks in both types of cell to ascertain the mechanism by which cancer is affected and healthy cells are not.
Biochemical Property Prediction:
BioPPsy is a package to predict clinically relevant properties of small molecules from those molecules from which such properties are already known. It works by training the chosen model with a given set of molecules and then using the trained model to predict the desired property, typically membrane permeability, to molecules for which this is unknown. When I took over this project it only had linear models available and was not yet on a public repository. Apart from some debugging and improving its structure in places, I have since added Partial Linear Squares, Neural Networks and Support Vector Regression (these last two by implementing the weka package). Just as importantly, this has been done in such a way that anyone can code their own algorithm and simply add it to the project. BioPPsy may be downloaded from my sourceforge repository.
Prediction of molecules in solution:
I have greatly improved the running time for some fortran code for predicting the properties of molecules in solution by parallelizing parts of the code using CUDA, and otherwise upgrading the code to modern coding standards.
Interpretation of Quantum Mechanics:
A mechanistic interpretation of quantum measurement is that mixed states generate off-diagonal contributions to the energy of a measuring device, prompting self-interaction and spontaneous violation of unitarity leading to wave-function collapse.