|
Abstract:
|
Although a protein can be characterized by its amino acid sequence first, most proteins fold into 3-D structures that determine their functions. After many successful attempts of sequencing, the focus is on structural bioinformatics where techniques are to be developed for predicting the secondary, tertiary, and quaternary structures of protein molecules from the sequence data. Several decades of research failed to produce any reliable prediction method, as experimental methods such as X-ray crystallography and NMR spectroscopy are slow and expensive and computational searches in the space of all possible 3-D configurations is impractical. Recently, breakthroughs have been achieved using statistical and machine learning approaches. After a quick review of the background needed, we will discuss the challenges involved and give a brief tour of these new approaches.
|
