The future of structural bioinformatics in the post-genomic era

Abstract

Protein similarity searches are still based on primary sequence data searching rather than structural information. There are large databases available nowadays, which contain huge amounts of structural and sequence information. However, in many cases, primary sequence comparisons do not cover sequence space adequately, creating an urgent need for a new structure-based similarity search approach. It is universally known that a protein's tertiary structure is more conserved than its primary structure. There are numerous cases of extremely low sequence similarity between homologous viral proteins (~10% sequence identity) that nevertheless have the same function, belong to the same viral family and, more strikingly, when their structures have been determined by X-ray crystallography, have shown the same structural features. Taken together, we propose that the novel structural similarity approach proposed here, in conjunction with classic sequence similarity BLAST searches, could yield results of great interest to science.