Show Abstract ▼

Abstract: The work of this thesis focuses on the development of novel computational methods for the determination of protein structures through macromolecular X-ray crystallography (MX). The main focus of Chapters 2 and 3 is the development of alternative molecular replacement (MR) approaches in cases where no related structure is available as a search model. In Chapter 2, the performance of a library of helical ensembles created by clustering helical segments is explored. A 30% increase in the number of solutions obtained using these search models was observed when compared with the performance recorded for single-model ideal helices. In Chapter 3, SWAMP is presented: a novel pipeline for the solution of structures of transmembrane proteins. SWAMP includes a library of ensembles built by clustering commonly observed packings of transmembrane helical pairs in close contact. The search models in this library are then ranked based on the similarity between their observed residue contacts and the contacts predicted for the unknown structure. Results show that SWAMP is capable of detecting valid search models originating from unrelated solved structures solely exploiting this contact information. In Chapter 4, the main focus of the work presented remains MR, particularly, the importance of experimental data collection and the quality of the obtained diffraction data. Specifically, the relation between data completeness, the distribution of missing reflections and the quality of the maps obtained through MR is studied. For this purpose, a set of new metrics for the distribution of missing reflections in the reciprocal lattice are proposed, and a large-scale study to assess the effects of data incompleteness on MR outcome is carried out. Results revealed low resolution completeness as a major factor affecting the quality of the maps obtained through MR, highlighting the importance of low resolution reflections in the process of MR. Overall data completeness, signal-to-noise ratio and search model quality were also other factors observed to determine, in conjunction MR outcome. In Chapter 5, new metrics for model validation are presented. These metrics are based on the availability of accurate inter-residue distance predictions, which are compared with the distances observed in the emerging model. These metrics were fed into a support vector machine classifier that was trained to detect model errors based on historical data from the EM Validation Challenges. Further analysis of the possible register errors is done by performing an alignment of the predicted contact map and the map inferred from the contacts observed in the model. Regions of the model where the maximum contact overlap is achieved through a sequence register different to that observed in the model are flagged and the optimal sequence register can then be used to fix the register error. Results suggest that both the detection of model errors and the correction of sequence register errors is possible, even in challenging cases, through the use of the trained classifier in conjunction with the contact map alignment. The approach, implemented in ConKit, thus provides a new tool for protein structure validation that is orthogonal to existing methods. Lastly in Chapter 6 ConPlot is presented: a web-based application which uses the typically empty space near the residue contact map diagonal to display multiple coloured tracks representing other sequence-based predictions. The integration of these different sources of information enables researchers to easily analyse a variety of data simultaneously.