Introduction
The UK collaboration “Deep and Frequent Phenotyping Study” has been running for 3 years and has generated a wealth of new data on Alzheimer's disease. The Alan Turing Institute and The Translational Neuroscience and Dementia Research Group, within the Department of Psychiatry, University of Oxford, are holding a 3 day datathon aimed at bringing cutting edge data science to this study and to the challenges of Alzheimer's research.
We are looking for 25 early career data scientists, as well as more experienced colleagues to join this workshop. The participants are expected to have technical and theoretical data science expertise, practical experience with Python / R software with standard data science and statistical packages. We would expect participants to demonstrate that they have collaborative skills and be able to convey findings to others in the workshop from outside of their specialty.
Spaces for this workshop are limited, applications are now closed.
If you have any questions, please do not hesitate to contact the organisers at [email protected]
About the event
The goal is to use a pilot data set to frame analysis that can be applied to the full data set of more than 200 subjects when it becomes available. This data workshop on the pilot Deep and Frequent Phenotyping Study data (DFP) is a unique opportunity to bring researchers involved with, or interested in working with rich and multimodal data on neurodegeneration.
The aim is to develop an engaged and connected community of scientists from diverse data science and clinical backgrounds. It will also be to initiate and establish collaboration in translational neuroscience and dementia, with the ambition of nurturing interest from the community, focusing it with the plot data, and marshalling it into effective and leading research in the field of data-driven Alzheimer's and Dementia research based on the full data set as it emerges.
During a pilot phase of the Deep and Frequent Phenotyping the following data modality have been collected from 22 participants all diagnosed with Alzheimer’s disease (AD) over the period of 6 months. The data types include: time-series data, raw and preprocessed brain and retinal images, structured numerical tables of surveys and clinical assessments, wearable triaxial sensor data, electromagnetic brain signal data (MEG/EEG) and biomarker data. More information can be found here.
The scientific objectives are to understand rich and multimodal DFP data and to allow the development of data science pipelines; with priorities being:
- to identify markers for progression of Alzheimer’s disease.
- multi-modal and unsupervised learning in neurodegeneration
The workshop will encourage participants to share/use novel and advanced mathematical methods in data analysis and machine to learn efficiently from complex real-world data.
Senior investigators from the DFP study will present the data and challenge the participants with specific problems. Participants will then split into groups to work collaboratively on the challenges.