Dr Frank DondelingerLecturer in Biostatistics
I am a Lecturer in Biostatistics at Lancaster University in the CHICAS (Combining Health Information, Computation And Statistics) group of the Lancaster Medical School. My research interests include Bayesian models for drug and treatment response, information sharing, transfer learning and group mapping for transferring knowledge from in vitro to in vivo datasets, network reconstruction from time-series and interventional data, efficient inference in complex high-dimensional Bayesian models, and parameter inference in models of biological systems.
My research focuses on applying machine learning techniques to practical problems in biological and biomedical research. I am particularly interested in problems involving network inference, such as inferring gene regulatory networks from gene expression data. I have applied Bayesian modelling techniques to this problem, including stationary and dynamic Bayesian networks. I have also applied sparse linear regression techniques for network inference. I have developed and tested new priors for information sharing in situations where the structure of the network can change.
Recently, I have been working on methods for parameter inference in ODE models of biological systems. Through my work at The Netherlands Cancer Research Institute, I have learned about the biology of cancer and have become interested in using hierarchical regression models for prediction on heterogeneous datasets, specifically in the context of drug response prediction.
PhD Supervision Interests
Please contact me if you are interested in pursuing a PhD in machine learning or statistical modelling for molecular biomedicine. I am currently taking applications for a funded 3-year PhD project (co-supervised with Professor Peter Diggle and Dr James Hensman) entitled: "Design and analysis for longitudinal animal studies with high-dimensional outcomes." Two other potential PhD topics without attached funding that I am currently offering are: 1) "Data-driven cancer subtype discovery using high-dimensional integrative clustering" and 2) "Machine Learning for Drug Selection in Personalized Cancer Medicine".
Molecular heterogeneity at the network level: high-dimensional testing, clustering and a TCGA case study
Staedler, N., Dondelinger, F., Hill, S., Akbani, R., Lu, Y., Mills, G., Mukherjee, S. 15/09/2017 In: Bioinformatics. 33, 18, p. 2890-2896. 7 p.