Sequence assignment for low-resolution modelling of protein crystal structures.

Grzegorz Chojnowski,Victor S Lamzin,Joana Pereira

doi:10.1107/s2059798319009392

Abstract

The performance of automated model building in crystal structure determination usually decreases with the resolution of the experimental data, and may result in fragmented models and incorrect side-chain assignment. Presented here are new methods for machine-learning-based docking of main-chain fragments to the sequence and for their sequence-independent connection using a dedicated library of protein fragments. The combined use of these new methods noticeably increases sequence coverage and reduces fragmentation of the protein models automatically built with ARP/wARP.

Highlights

Model building is a key step in crystallographic structure determination
The combined use of these new methods noticeably increases sequence coverage and reduces fragmentation of the protein models automatically built with ARP/wARP
We present two new methods incorporated into the ARP/wARP software that address the protein side-chain assignment in crystallographic structure determination at low resolution, especially when only incomplete, fragmented and often main-chain-only models are available

Summary

Introduction

Model building is a key step in crystallographic structure determination. When the resolution of the X-ray diffraction data is better than 3.0 Aand the initial map is of reasonable quality, model building can often be accomplished straightforwardly using automated approaches. In difficult cases automated approaches may not fully succeed, but may still help to improve the electron-density maps to a level that enables unambiguous manual interpretation. This is important owing to recent advancements in molecularreplacement and experimental phasing pipelines [examples include BALBES (Long et al, 2008), MrBUMP (Keegan & Winn, 2007), MORDA (Vagin & Lebedev, 2015), ARCIMBOLDO (Sammito et al, 2014) and Auto-Rickshaw (Panjikar et al, 2009)], where automated model building is often used for the evaluation of plausible solutions (Ha & Boggon, 2018). In the presence of significant phase error and/or with limited resolution of the experimental data (worse than 2.5 A ) the model-building task remains challenging and, with the need for manual intervention, becomes time-consuming even for an experienced crystallographer

Methods

Results

Conclusion