AI4Green
Research Topics
Ab Initio Calculations of the Electronic Excited States of Molecules,
Electronic Structure and Circular Dichroism
of Proteins, Protein Folding
and Evolution, Bioinformatics, Computer-Aided Drug Design, Drug Resistance.
Please follow the links to publications on the respective topic.
Research In Progress
 |
Optical properties of proteins, such as circular dichroism, provide useful experimental probes of protein folding. Such experiments can be performed on the nanosecond time scale and allow the observation of some of the earliest events in protein folding. To enhance the interpretation of these important experiments, we are improving theoretical methods for calculating the circular dichroism of proteins. These calculations are being coupled with all-atom molecular dynamics simulations of protein folding to provide a detailed connection between protein conformation and observable spectroscopic properties. |
Electronic Excited State Calculations
 |
Of the many roles that solvent plays, its influence on molecular electronic structure is a particularly challenging phenomenon to study. We are exploring a combination of implicit continuum models of solvent and explicit solvent molecules. Recent ab initio methods are used compute the electronic structure of small molecules that are important models of chromophoric groups in proteins. |
Protein Folding and Evolution/Bioinformatics
 |
To study the distinct influences of structure and function of evolution, we have developed a highly simplified computational model of proteins with binding pockets, called functional model proteins. The configuration of the polypeptide chain is confined to a lattice, reducing the number of possible structures from an astronomical number to a computationally manageable number. The number of different types of amino acid is also considerably reduced. We have explored the fitness or evolutionary landscapes, as characterised by the size and distribution of homologous families and by the complexity of the inter-relatedness of the functional model proteins. In the context of bioinformatics, understanding the evolutionary landscapes of these models may help rationalise the deluge of genomic sequence data. We are currently extending the models to be closer to real proteins. |
 |
We are working on the application of non-parametric statistical methods in the development of quantitative structure-activity relationships (QSAR) used to rationalise the biological activity of molecules and to aid the design of further compounds in drug discovery efforts. The advent of combinatorial chemistry has made many thousands of compounds accessible and computational methods offer a means for a fuller exploitation of combinatorial chemistry approaches to drug design. |