The subject of bioinformatics covers all the areas in which computers and computational/mathematical techniques are applied in storing, managing, mining and interpreting the vast and ever- increasing amount of sequence and structure data arising from the work of experimental molecular biologists, in particular nowadays from the various genome sequencing projects. The aim of this computational and analytical work is to bring to light the underlying principles which are hidden in these mountains of data, and to give us a better understanding of how biological systems operate at the molecular level: how proteins find their correct 3-dimensional structures, how proteins interact, how complex biological systems may have evolved from much simpler precursors, how genes are regulated in growth, development and ageing.

The work which I have been involved in is specifically related to using neural network algorithms and other mathematical techniques - for example Fourier and wavelet transformation methods - to try to predict how proteins fold and how they associate into more complex multimeric molecules. It has been carried out in collaboration with Prof Janet M Thornton and members of her Molecular Structure Group (formerly here at UCL, now largely relocated to Hinxton following Prof Thornton's move to the EBI in January 2002), in particular the following people:

Dr Adrian J Shepherd (formerly UCL Biochemistry, now Birkbeck Crystallography): use of neural networks to predict the presence of local folding motifs and to classify types of protein folds using a Fourier-based preprocessing method

Dr Kevin Murray (formerly UCL Biochemistry, now EBI): use of wavelet transforms in fold classification and to detect structural repeats in proteins

Dr Hannes Postingl (EBI) and Dr Thomas Kabir (former joint UCL-CS/Biochemistry PhD student): systematic characterisation of protein-protein interfaces in trimers, tetramers and higher multimers; use of neural networks to predict the locations of such interfaces.