MRC DTP 4 Year PhD Programme: Stratifying patients by drug response in a large diabetic cohort using deep learning

Aims and objectives. The project aims to stratify diabetic patients by their response to drugs using artificial intelligence (DL) techniques, in line with the increasing interest for Big Data and AI of the School of Medicine.

Experience and environment. Pearson’s group has extensive and internationally visible experience of diabetes research and response to drugs. Pearson leads the €46M IMI-DIRECT project on stratification in Type 2 diabetes and is Strand 2 lead on the £2.7M MRC funded MASTERMIND project. Trucco’s group (CVIP/VAMPIRE) has extensive experience of DL applied to medical image and data analysis, and runs its own GPU resources (~10 NVIDIA cards). GoDARTS, already used in various collaborative computing-medicine projects, has comprehensive consent for research. The student will join the friendly and active CVIP/VAMPIRE laboratory and be attached to the student cohort of the NIHR Precision Medicine for Diabetes (Pearson, Trucco investigators) for effective interdisciplinary integration.
Materials. We shall rely on GoDARTS, a large, cross-linked bioresource grown in the medical school including more than 9,000 diabetic patients and 8,000 controls (http://diabetesgenetics.dundee.ac.uk/) . In additional to a variety of personal, lifestyle and clinical measurements and information, GoDARTS includes treatment history (which drugs, when prescribed, patient progress), outcomes, and full genetic profiles. This is excellent ground for data mining with artificial intelligence techniques.

Training and work plan. (1) Initial induction: familiarization with interdisciplinary computing-medicine group; literature review (DL and drug response; related topics); training in machine / DL, clinical statistics, clinical context; short project (~2-3 months) leading to conference paper. (2) Identification, implementation, test of DL architectures for target problem; investigation on transfer learning and pre-trained large networks. (3) Definition of final specific problem expected after first year.

References.

R Annunziata, E Trucco: Accelerating Convolutional Sparse Coding for Curvilinear Structures Segmentation by Refining SCIRD-TS Filter Banks. IEEE Trans on Medical Imaging, vol 35 no 11, Nov 2016, pp 2381-2392.

A E Fetit, S Manivannan, S McGrory, L Ballerini, A J Doney, T MacGillivray, I J Deary, J M Wardlaw, F Doubal, G J McKay, S J McKenna, E Trucco. Retinal Biomarker Discovery for Dementia in an Elderly Diabetic Population. Proc MICCAI International Workshop OMIA-4, Springer, 2017.

Zhou, K, Pedersen, HK, Dawed, AY, Pearson, ER 2016, Pharmacogenomics in diabetes mellitus: insights into drug action and drug discovery. Nature Reviews Endocrinology, vol 12, no. 6, pp. 337-46. DOI: 10.1038/nrendo.2016.51 .

H Shen, R Wang, J Zhang, S McKenna: Boundary-aware Fully Convolutional Network for Brain Tumor Segmentation. Proc Intern Conf on Medical Image Computing and Computer Assisted Intervention (MICCAI), Quebec City, Canada, 2017. DOI: 10.1007/978-3-319-66185-8_49 .