Diploma projects

(last update: 5 March 2009)

Quantum mechanical modelling of cancer diagnostics agents

The recent progress in tumour diagnostics has resulted in a shift of emphasis in clinical oncology from tumour treatment to early tumour detection. MRI (magnetic resonance imaging) methods are particularly important in this respect, as they allow rapid and non-invasive tumour detection, provided that the so-called 'contrast' between healthy and cancerous tissue can be made sufficiently large. The MRI contrast depends on magnetic properties of the tissue and a number of molecules are known that show an abrupt change in those properties depending on the in vivo chemical environment (pH, oxidation potential, etc). Because tumour tissue is frequently different (more acidic and more oxidative) than the healthy tissue, high MRI contrast can in many cases be achieved and the early tumour identification performed successfully.

The development of such contrast agents is crucially important, and relies not only on the synthetic methodology used in their production, but also on the detailed understanding of their magnetic properties. Prediction and rationalization of the latter is frequently a relativistic quantum mechanical problem, requiring state-of-the-art solvated DFT geometries, GIAO calculations of magnetic shielding and magnetic susceptibility tensors with full account of spin-orbit coupling and conformational exchange. In many cases, computational pre-screening of contrast agent candidates can save months of expensive synthetic work and give unique insights into the kinds of structures that produce the desired magnetochemical response.

Computational magnetochemistry and quantum mechanical modelling of reaction dynamics

Chemical shifts, hyperfine couplings, g-factors and other magnetic properties are very difficult to measure experimentally for large, short-lived or quickly relaxing molecules, but they can be computed with very high accuracy using modern DFT (density functional theory) techniques. You will receive training in the use of the density functional theory and perform calculations of magnetic coupling tensors in a number of organic radicals and proteins, which are encountered in magnetosensitive chemical reactions and protein folding research.

The reactions in question belong to the exciting class of the so-called 'compass reactions' which change their yield in response to a change in applied magnetic field. They were recently shown to provide migratory birds with an additioinal navigation mechanism -- the ability to directly sense the magnitude and inclination of the Earth's magnetic field. They are also considered in the context of the (thus far unidentified) possible health effects of extensive mobile phone use.

Using graphics processors to accelerate spin dynamics simulations

This project will explore one of the exciting innovations in modern high-performance computing -- conventional data processing and quantum mechanical simulations using the enormous computing power of modern graphics cards. The performance which can be achieved for some applications is stunning: each of the 128 cores on the Nvidia 8800GTX card has a floating point capability comparable to each of the cores in an Intel Core 2 Duo CPU. You will receive training in scientific computing and programming and perform a number of benchmark and production simulations of spin systems commonly encountered in chemical and biological magnetic resonance.

Highly efficient spin dynamics simulation algorithms

This is a theoretical project dealing with an old and important problem in chemical and biomedical magnetic resonance -- that of the exponential scaling of simulation complexity with the number of particles in the system. Put simply, exponential scaling means that systems of more than about 10 coupled spins are currently computationally intractable -- a major limitation, particularly in biomolecular NMR spectroscopy. We have recently developed a polynomially scaling algorithm, which allows much bigger systems to be treated. You will receive training in spin dynamics, NMR and scientific computing, develop and implement the high-efficiency algorithms for one- and multidimensional NMR experiments.

Please contact Dr. Ilya Kuprov for further information.