Accounting for correlated errors with maximum likelihood in crystallography and cryo-EM

Projektinformationen

LikelyStructures

ID Finanzhilfevereinbarung: 790122

DOI

10.3030/790122

Projekt abgeschlossen

EK-Unterschriftsdatum 19 März 2018

Startdatum 1 Juni 2018

Enddatum 31 Mai 2020

Finanziert unter

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Gesamtkosten

€ 183 454,80

EU-Beitrag

€ 183 454,80

Koordiniert durch

THE CHANCELLOR MASTERS AND SCHOLARS OF THE UNIVERSITY OF CAMBRIDGE
United Kingdom

Periodic Reporting for period 1 - LikelyStructures (Accounting for correlated errors with maximum likelihood in crystallography and cryo-EM)

Berichtszeitraum: 2018-06-01 bis 2020-05-31

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Macromolecular Crystallography is a gold standard experimental technique to determine the structures of proteins and other macromolecules. As only partial information is directly measured in the experiment, it is usually necessary to use computational methods to derive the missing information. Molecular Replacement (MR) bootstraps the final structure from the initial partial information given by other molecules that share a certain degree of similarity with the unknown structure. In particular, the Phaser MR software uses a statistical approach based on maximum likelihood that is able to exploit even low signal from such remote models, and it is today the most-used software worldwide to accomplish such a task. Over the years the mathematical foundations of the software have been strengthened and the algorithms have become more sophisticated, but limitations remain because of uncertainties in the quality of models and in assumptions about the data themselves. In this project, we tried to address some of those uncertainties by building an automatic, easy to use, graphical pipeline that will prepare both the data and the models prior to and specifically for the molecular replacement task. The findings in this procedure can be exploited and applied also to emergent techniques such as Cryo-EM and EM-Tomography. Our main goal is to build a unified framework that can tackle structural biology problems from different points of view and exploiting prior information combined with machine learning approaches.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

The results of the project are all integrated in the software Voyager that will soon be released, initially in the crystallographic community, and then extended to the Cryo-EM and EM-Tomography community. Voyager combines Maximum Likelihood approaches to analyse and correct experimental data and then to perform molecular replacement. The majority of macromolecular structures are solved by this technique, though its application might be difficult for non-expert users especially in the presence of complex molecular structure with only poor homology models available. Voyager automates and rationalises the preparation of both experimental and model data improving the chances of success, providing clear and reproducible results that will push the boundaries of current molecular replacement, and provide better optimised initial phases for model building and refinement.

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

Voyager will contribute significantly to the process of structure determination in structural biology, simplifying operations and pushing the boundaries toward even larger complexes with large domain movement, coping better with low-resolution data that present pathologies. The software will impact positively not only on academic research but also on the pharmaceutical industries where the quality of solved structures and the precise understanding of their quaternary conformations is fundamental to plan and deliver high-quality drugs.