Needle in the haystack: structure-based toxin discovery

In the current data-rich era, making the leap from sequence data to knowledge is a task that requires an elegant bioinformatics toolset to pinpoint pressing research questions. Therefore, a strategy to expand important protein-family knowledge is required, particularly in cases in which primary sequence identity is low but structural conservation is high. For example, the mono-ADP-ribosylating toxins fit these criteria and several approaches have been used to accelerate the discovery of new family members. The strategy evolved from conduction of PSI–BLAST searches through to the combination of secondary-structure prediction with pattern-based searches. However, a newly developed tactic, in which fold recognition dominates, reduces reliance on sequence similarity and advances scientists toward a true structure-based protein-family expansion methodology.