Designing more efficient compound libraries for crystallographic fragment screening - doctoral thesis

Aug 3, 2026

Upload by Maeve

Fragment-based drug design explores protein binding sites and supports the development of bespoke, specific, high-affinity drugs. Small fragment size enables an efficient search of chemical space, increasing hit rate and reducing the number of fragments required versus screening larger molecules, thereby lowering cost and time in structure-based drug design pipelines. In established campaigns, incomplete binding-site characterisation limits the progression of diverse lead compounds. The research uses retrospective analysis of XChem fragment screening data to curate reliable datasets, improve storage and automation of electron-density maps, select functionally diverse fragments with maximised information gain, and propose target-specific strategies using machine learning, point-cloud clustering, and pocket-similarity and protein representation tools.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages234

Reading Time9 h 50 min

AccessFree to download

Abstract

Fragment-based drug design advantages and limitations

Hypothesis and aims for maximising binding-site information

Survey and curation of XChem data quality

Cloud-based storage and automated electron-density curation

Functionally diverse fragment selection and ranking

Target-specific library design with binding-site prediction and pocket similarity

Designing more efficient compound libraries for crystallographic fragment screening Anna Carbery Lady Margaret Hall University of Oxford A thesis submitted for the degree of Doctor of Philosophy Michaelmas 2023 Acknowledgements First, I’d like to thank Charlotte, who has now seen me through two degrees and a pandemic. Her consistent guidance, support and patience are hugely appreciated. I’m also immensely grateful to Rachael, who has not only been an invaluable sounding board and a source of a great depth of knowledge, but whose confidence in me has helped me through the highs and lows of this DPhil. Thanks also go to Frank, who ensured my research would always be of use to the wider community. I would like to thank all OPIGlets past and present for their constant energy and enthusiasm for all things science and otherwise. The long lunches and gossip sessions will be sorely missed, and I hope to make any future workplace as OPIG-like as possible. Special shoutouts go to Lucy, Kate, Gemma, Alissa, Maranga, Olly, Guy and ´Isak, whose camaraderie has made the final year of this DPhil the best. Outside of OPIG, I’m grateful to have close friends who have helped me throughout my eight years in Oxford. I wouldn’t have made it to the start of this degree without Jess, Helen, Immie and Jade, and the DPhil years have been made all the better by getting out of Oxford to visit them all. Extra special thanks goto Alice and Agnes for being amazing flatmates and friends who made Oxford really feel like home. Coaching for LMHBC M1 has been a highlight of my time in Oxford, and I’d like to thank all the generations of crews that I coached for their enthusiastic attempts to put my unintelligible rowing analogies into practice. In particular, thanks go to Harry, Owen, Ben and Will, who made me feel at home from the very beginning. I am hugely grateful to my family for all their love and support. To John, who was always there to pick up the pieces. To my sister Helen, whose many daily phone calls meant that she was always up to speed. And to my parents, Tony and Maria, who inspired my love of science and empowered me to push myself. Abstract Fragment-based drug design aims to explore the binding sites of protein targets and provide information towards bespoke, specific, high-affinity drugs. The small size of fragments facilitates an efficient search of chemical space; only a few efficient interactions need to be made with the target for binding to occur, thus improving the hit rate and reducing the number of fragments that need to be screened in comparison to traditional screening of larger molecules. This approach aims to facilitate a much more efficient structure-based drug design pipeline, decreasing the cost and time taken per drug. Because of the small size of fragments, in established fragment screening campaigns, even with a good hit rate, the binding site is often not fully characterised. This reduces the number of diverse lead compounds that can be progressed from the initial fragment hits, and makes it more expensive and difficult to develop a viable drug. We hypothesise that this occurs because fragments within screening libraries are traditionally selected for diverse coverage of a broad chemical space, and do not consider functional information such as coverage of potential protein-ligand interactions. This research aims to use retrospective analysis of fragment screening data to develop methods for selecting fragment libraries that will maximise the information obtained from a target’s binding site. We first describe a survey of data availability and quality for XChem fragment screens. The curation procedure for fragment screening data presents many challenges, such as storage of raw crystallographic data and variability within the analysis pipelines. We found that the annotations applied to historic fragment screening data were unreliable due to the complexities of building and validating ligand models. A

FAQs

What problem does the thesis address in fragment screening campaigns?

Even with a good hit rate, fragment screening often leaves the binding site incompletely characterised, which reduces the number of diverse lead compounds and makes viable drug development more expensive and difficult.

How does the thesis improve the use of XChem fragment screening data?

It surveys data availability and quality, identifies unreliable annotations and difficulties assembling complete datasets, then proposes cloud-based storage plus automated backup and curation of crystallographic electron-density maps to enable more consistent automated modelling and validation.

What method is proposed for selecting fragments from a larger library?

The thesis ranks fragments by the number of novel interactions they make with historic protein targets, using positive and negative data extracted from XChem to show improved recovery of information on unseen targets compared with traditional selection.

Aug 3, 2026

Upload by Maeve