Fragon: rapid high-resolution structure determination from ideal protein fragments

Abstract

Correctly positioning ideal protein fragments by molecular replacement presents an attractive method for obtaining preliminary phases when no template structure for molecular replacement is available. This has been exploited in several existing pipelines. This paper presents a new pipeline, namedFragon, in which fragments (ideal α-helices or β-strands) are placed usingPhaserand the phases calculated from these coordinates are then improved by the density-modification methods provided byACORN. The reliable scoring algorithm provided byACORNidentifies success. In these cases, the resulting phases are usually of sufficient quality to enable automated model building of the entire structure.Fragonwas evaluated against two test sets comprising mixed α/β folds and all-β folds at resolutions between 1.0 and 1.7 Å. Success rates of 61% for the mixed α/β test set and 30% for the all-β test set were achieved. In almost 70% of successful runs, fragment placement and density modification took less than 30 min on relatively modest four-core desktop computers. In all successful runs the best set of phases enabled automated model building withARP/wARPto complete the structure.