This thesis presents an account of two important facets of glycobioinformatics, comprising database development and molecular modeling of 3D structures of carbohydrates alongside the simulation of protein-carbohydrate interactions. Classical molecular modeling techniques were used to reconstruct 3D polysaccharide structures from experimentally determined atomic coordinates, or known starting points about their structures were used as guidelines to model them. A genetic algorithm search was employed as a high-throughput technique to characterize low energy conformers of bioactive oligosaccharides. The data generated were organized into two open-access relational databases, namely, PolySac3DB and BiOligo, for use by the scientific community. The validation of the molecular techniques used were performed using solution phase NMR experiments on four entero aggregative pathogenic E. coli strains, and were found to be robust and realistic. Further, the impact of the presentation of human fucosylated oligosaccharide epitopes to lectins from opportunistic gram negative bacteria, was investigated in a screening study using molecular docking studies, which could help in evaluating the feasibility of using automated docking procedures in such instances as well as deciphering binding data from glycan array experiments and also correlated to isothermal calorimetry data. On comparison with high-resolution experimental crystal complexes, automated docking was found to delineate the present level of applicability, while emphasizing the need of constant monitoring and possible filtering of the results obtained. Finally, a review of the present status of the computational aspects of protein-carbohydrate interaction studies is discussed in the perspectives of using molecular modeling and simulation studies to probe this aspect of molecular and structural glycobiology.
