Frank Dondelinger
Research Student
Biomathematics & Statistics Scotland
JCMB, The King's Buildings,
EDINBURGH, EH9 3JZ, Scotland, UK.
Tel: +44 (0)131 650 7536
Fax: +44 (0)131 650 4901
Email: Frank Dondelinger
My Research
Interests
- Network inference: Bayesian networks, graphical models, sparse regression
- MCMC and other sampling approaches
- Dynamic models and ODE systems
- Applications of these techniques to ecology and systems biology
Publications
- F. Dondelinger, S. Lèbre and D. Husmeier, "Non-homogeneous dynamic Bayesian networks with Bayesian regularization", Machine Learning (2012) [Under Review]. [Software]
- F. Dondelinger, D. Husmeier and S. Lèbre, "Dynamic Bayesian networks in molecular plant science: inferring gene regulatory networks from multiple gene expression time series", Euphytica (2011). [LINK]
- D. Husmeier, F. Dondelinger and S. Lèbre, "Inter-time segment information sharing for non-homogeneous dynamic Bayesian networks", NIPS (2010). [PDF] [Software]
- F. Dondelinger, D. Husmeier and S. Lèbre, "Heterogeneous Continuous Dynamic Bayesian Networks with Flexible Structure and Inter-Time Segment Information Sharing", ICML (2010). [PDF] [Software]
- A. Faisal, F. Dondelinger, C. Beale and D. Husmeier, "Inferring species interaction networks from species abundance data: A comparative evaluation of various statistical and machine learning methods," Ecological Informatics (2010). [LINK] [Software and Supplementary Material]
- F. Dondelinger, "Inferring Ecological Networks From Species Abundance Data", MSc Thesis (2008) [PDF]
Ecological Networks
In this work, we inferred ecological networks of species interactions from species abundance data using multiple machine learning methods. This was joint work with Ali Faisal, Dirk Husmeier and Colin Beale, and was published in Ecological Informatics.
Below you can find the Matlab software for the network inference methods, as well as an a priori species interaction network of warbler species compiled from the literature by Colin Beale that we used for evaluation purposes and the supplementary material for our paper.
- Simulation Software - This Matlab software simulates ecological interactions between species in a food web. Based on code by Jon Yearsley.
- Bayesian Network Inference - This Matlab software uses a Markov Chain Monte Carlo algorithm to infer the structure of a Bayesian network from ecological population data.
- A Priori Network - This Excel file lists species interactions among warblers that have been reported in the literature, along with our confidence in the presence of the interaction and the relevant references.
- Supplementary Material - Extra information relating to our paper on inferring ecological networks.
Gene Regulatory Networks
This work concerns the problem of inferring gene regulatory networks from gene expression data. In particular, we are looking at scenarios where the gene regulatory network may change over time. We developed and applied heterogeneous dynamic Bayesian network, where a changepoint model allows the network structure to change.
The software for doing inference in this model has been written in R, and will be available on CRAN once we have finished documenting it. A sparsely-documented version of the software is available here.
Brief CV
- 2003 - 2007: BSc. in Artificial Intelligence and Computer Science, University of Edinburgh
- 2007 - 2008: MSc. in Artificial Intelligence (Specialisations Learning from Data and Bioinformatics), University of Edinburgh
- 2008 - Present: PhD in Machine Learning Approaches for Network Inference in Ecology and Molecular Biology, BioSS
Other Interests
I contribute as an author and editor to the Edinburgh University Science magazine (EUSci).
