Statistically robust methods for the integration and analysis of X-ray diffraction data from pixel array detectors

New challenges in structural biology are driving the development of new technology, such as photon counting pixel array detectors, and new modes of data collection, such as serial synchrotron crystallography (SSX), in macromolecular crystallography at synchrotrons. This in turn is creating the need for better algorithms and software to extract the maximum amount of information from the diffraction data. The aim of this project is to develop statistically robust methods for the integration and analysis of X-ray diffraction data to address these challenges. A method for estimating the background under each reflection during integration that is robust in the presence of pixel outliers is presented. This uses a generalised linear model (GLM) approach that is more appropriate for use with Poisson distributed data than traditional approaches to pixel outlier handling in integration programs. The algorithm is most applicable to data with a very low background level where assumptions of a normal distribution are no longer valid as an approximation to the Poisson distribution. A second algorithm for modelling the background for each Bragg reflection in a series of X-ray diffraction images containing Debye-Scherrer diffraction from ice in the sample is also presented. This method involves the use of a global background model which is generated from the complete X-ray diffraction dataset. Fitting of this model to the background pixels is then done for each reflection independently. Finally, a model for the observed reflection profiles is described for the purpose of improving the refinement of the crystal unit cell and orientation for still image diffraction data collected at synchrotrons. This model consists of two components: a Normal distribution is used to describe the distribution of wavelengths and a Multivariate Normal distribution (MVN) is used to describe the distribution of reciprocal lattice vectors for each reflection; this allows non-isotropic spot shapes to be easily described. The parameters of the model are estimated from the data via a simple maximum likelihood algorithm. The algorithms are incorporated into the DIALS integration package.

Name dc:type.qualificationname

Doctor of Philosophy (PhD)

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

University of Cambridge

Year dc:date.issued

2020

Author dc:creator

• Parkhurst, James Michael

Advisors dc:contributor.advisor

• Murshudov, Garib
• Evans, Gwyndaf

• Macromolcules
• Crystallography
• Algorithms
• Integration

Licence

• https://www.repository.cam.ac.uk/bitstreams/7ea03b4f-cbb0-4a73-ac07-c7f77778793b/download
• https://www.rioxx.net/licenses/all-rights-reserved/

Language dc:language

en

DOI dc:identifier.doi

https://doi.org/10.17863/CAM.46755

OAI identifier oai:identifier

oai:www.repository.cam.ac.uk:1810/299685