Research Enrichment – Open Research: people we've funded
Professor Alastair Denniston
University Hospitals Birmingham NHS Foundation Trust
Measuring uveitis activity using spectral-domain optical coherence tomography
The understanding of diseases of the retina, which can cause blindness, has been revolutionised by the light-based imaging technique optical coherence tomography (OCT). OCT allows visualisation of individual retinal layers which can allow diagnosis and treatment of disease. OCT segmentation and measurement of the retinal layers is restricted in standard commercial software, limiting researchers’ abilities to mine this wealth of data.
We will provide an open, accessible, cross-platform tool for the segmentation of OCT images. ReLayer is an online platform where OCT images can be uploaded and analysed. We have written a novel algorithm in Matlab which segments the scans and allows data extraction. We want to reduce processing time by translating the algorithm into a more efficient programming language. We will also extend platform compatibility to other popular OCT instruments enabling Relayer to process multiple images simultaneously. We will also develop accompanying online tutorials and support international outreach.
Our research will help to reduce processing times for OCT imaging which will help to speed up diagnosis of diseases of the retina.
Dr Meriem El Karoui
University of Edinburgh
DNA repair and genetic stability: elucidating the effects of cell physiology in Escherichia coli
One of the biggest problems of reusing data in open research is the quality of data description (metadata), which often lacks sufficient detail. The problem is acute in quantitative microscopy, which is the central method used in my project to measure DNA repair proteins in living bacteria. A carefully tuned sequence of image and data analysis steps is critical to track individual protein molecules. Each stage of analysis must be recorded and described to make the research reproducible.
We will develop a software platform that links the research tools that we use daily with existing, public and institutional repositories. It will automatically collect metadata at each step. It will assemble comprehensively described datasets and make them automatically available for open access at the university.
By developing a fully automated platform we will make open data sharing easier and more accurate. We will use the platform to share data with our collaborators which will improve working practices in other groups.
Dr Paul Flicek
Ensembl: a comprehensive reference resource for genomics
The Ensembl project provides free access to reference vertebrate genomes and associated data.
The aim of this open research activity is to increase Ensembl’s presence in low- and middle-income countries in Africa, Asia and South and Central America, adding two or three training workshops in these regions each year. This translates to training more than 200 scientists to use Ensembl and related free online tools for genomics in their research. A train-the-trainer element to the workshops will encourage information sharing, spreading bioinformatics skills even further. The skills learned in these workshops will allow scientists to use Ensembl open access data more effectively to contribute to their research, speeding up their analyses of genes and genomic variants. Participants will have access to presentations, walk-throughs, exercises and answers after the course, which they can share with colleagues.
This will improve the research output in those regions, which tend to focus on malaria, HIV and rice cultivation.
Dr Sally Lowell
University of Edinburgh
Differentiation competence of pluripotent cells
We will provide image analysis software that can extract information about how cells form functional and morphological networks in tissues or organoids based on 3D imaging data. These methods would be useful for understanding the cell to cell interactions that govern tissue formation, repair, and pathology. This software is built on a data-exploration framework that enables flexible interrogation of complex datasets. It is designed to be readily extensible and new segmentation methods from other groups can be incorporated into this single data-exploration system.
We will enable the most effective use of data and provide an environment for collaborative exploration of datasets by investing in a sustainable open-source format and providing a flexible database for data exploration.
Dr Ede Rancz
The Francis Crick Institute
Visuo-spatial processing in retrosplenial cortex
We aim to improve the methods available to investigate how neural circuits in the brain represent and process information using virtual environments to study the mouse brain. A major drawback of current virtual reality (VR) systems is the lack of vestibular stimulation, which effectively disables the brain’s internal compass.
We have developed a novel VR prototype which overcomes this limitation. With the help of a prototyping engineer, we will develop an open-source version for use by other laboratories. Open-source research equipment is economical, easy-to-build and can be easily adapted to new experiments. Curated open-hardware tools have allowed the unprecedented spread of cutting-edge experimental techniques. To reach as many users as possible, we will publish a methods paper and collaborate with the translation team at The Francis Crick Institute.
We will turn our prototype into a system which can be easily and economically incorporated into existing mouse VR set-ups or used alone by researchers around the world.