SUNY at Albany
June 19-23, 2001
Protein Structure Data Mining via the Kappa-Tau Framework. N- Glycosylation Sties in Human and Mammalian Proteins.
In the next years and decades we will be flooded with unpreceeded amounts of sequence and structural biomolecular data generated world wide and rapidly accessible via the internet. It is a major challenge to develop computational techniques powerful enough to cope with the volume and inherent complexity of the data, enabling us to recognise, evaluate, understand and predict.
We have recently, constructed (1,2), mathematically exact, highly efficient framework to encode protein structures based on fundamenal concepts of Differential and -Distance Geometry, addreviated the kappa-tau (curvature, torsion) framework. It can be used in large scale database screening for structural motifs, the classification of protein folds, structural alignments and the elucidation of sequence -structure relationships.
We here apply kappa-tau analysis to protein glycosylation, of subject of high tipical intrest(3), as a first step towards expanding the capabilities of available tools (4,5) for the automatic prediction of glycosylation sited from knowledge of the sequences. For this purpose we have quantitatively analyzed, compared and classified the local geometries of large, representative N-glycosylation site sets from human and mammalian proteins, including both esperimentally documented glycosylated sites (positives) as well as sites with sequences conforming to known patterns signalling glycosylation but not glycosylated (negatives).
First the results from this work concerning structural requirements for N-glycosylation and the effects of adjacent sequences and structures (context dependence) wil be presented.References and Footnotes
Dikeos Mario Soumpasis, Ramneek Gupta & Soeren Brunak
Center for Biological Sequence Analysis. Bldg 208 The Technical University of Denmark. 2800 Lyngby, Denmark