Supporting Binding-Sites Discovery via Iterative Database Processing

Abstract

Recognition of binding sites common to a set of protein structures is important for applications such as drug design. Common methods of binding-sites are based on heuristic algorithms that use summarized spatial data and superimposition techniques. However, computational operations generally do not store intermediate data for further calculation and information extraction. The current study presents an alternative approach to binding calculation by introducing a binary representation scheme for three dimensional molecule data and a fast iterative algorithm which obviates the need to calculate and resolve spatial transformations in the binding site extraction process. This is achieved by using relational database indexing methods and an efficient iterative model. This general-purpose iterative algorithm was tested for binding small molecules. The results show that the method can be applied efficiently for binding site extraction, and bio-information extraction. This binary representation improves performance by reducing processing time by 31% compared to typical representations.