Theoretical Systems Biology

I am currently a post-doc in the theoretical systems biology group at Imperial College London, performing research on the evolution of biological systems and methods for the reverse engineering of biological networks. My interests include the evolution of protein-protein interaction networks, inferring regulatory network structures from time-series data, and statistical methods for performing Bayesian inference on large data sets with high dimensional parameter spaces.

I completed my PhD in Systems Biology at Imperial College London in the theoretical Systems Biology group in 2010. Prior to this I obtained a BA in Computer Science from King’s College Cambridge and an MSc in Bioinformatics at Imperial College London.

Research activities

Statistical inference of biological networks and dynamical systems, including graphical models and Bayesian nonparametrics
Approximate Bayesian inference and Monte Carlo methods
Evolutionary modelling of biological systems

Current research

Many models of biological systems assume that the model incorporates all of the possible influences that may affect the behaviour of the system, but in practice this assumption is rarely justified. For example in the modelling of gene regulatory networks based on microarray timeseries data in may be the case that the presence or absence of some regulatory interactions are determined by factors not measured in the data and generally not included in simplistic models.

We can work around this problem by introducing a hidden state in the model that determines the structure of the model at each measured time point. To do so we apply the Hierarchical Dirichlet Process Hidden Markov Model (HDP-HMM), a method from Bayesian nonparametrics that allows us to model a hidden state sequence that can adapt in complexity to explain the observed data.

HDP-HMM

Then modelling the structure of the regulatory network as a Bayesian network, we can infer regulatory network structures that change over time based on gene expression microarray data. For example applying our method to gene expression data from the model organism Arabidopsis thaliana over a day night cycle we find two distinct regulatory network structures corresponding to the light and dark phases:

Athaliana inference

Recent oral/poster presentations

MASAMB 2013 (London) – Going beyond static networks (talk)
TBC 2012 (Jeju) – Graphical modeling of regulatory interactions in sporadic Inclusion Body Myositis (poster)
ISBA 2012 (Kyoto) – Interaction networks changing with time (poster)
BayesComp 2012 (Tokyo) – Sequential Monte Carlo inference of protein interaction network evolution via graph spectra (poster)
2011 UK-Japan Workshop on Microbial Systems Biology (Tsuruoka) – Interaction networks changing with time (talk)
MASAMB XXI (Vienna) – Sequential Monte Carlo samplers for biological network inference (talk)
BBSRC 5th Annual Systems Biology Grant Holders’ Workshop (London) – Regulatory network inference in Candida glabrata (talk)

Software

networks

SpecDist (see Graph spectral analysis of protein interaction network evolution.)

Sputnik stochastic petri net simulation software for biological models.

Publications

Graphical modelling of molecular networks underlying sporadic Inclusion Body Myositis.

Thorne T, Fratta P, Hanna M, Cortese A, Plagnol V, Fisher EM, Stumpf MP.

Molecular BioSystems 2013.

Model selection in systems and synthetic biology.

Kirk P*, Thorne T*, Stumpf MP. * Joint first authors.

Current Opinion in Biotechnology 2013.

Inference of Temporally Varying Bayesian Networks.

Thorne T, Stumpf MP.

Bioinformatics (2012) 28(24): 3298-3305

A parameter-free model discrimination criterion based on steady-state coplanarity.

Harrington HA, Ho KL, Thorne T, Stumpf MP.

Proc Natl Acad Sci U S A2012 vol. 109 no. 39 15746-15751

Considerate approaches to constructing summary statistics for ABC model selection.

Barnes CP, Filippi S, Stumpf MP, Thorne T.

Statistics and Computing 2012

Graph spectral analysis of protein interaction network evolution.

Thorne T, Stumpf MP.

Journal of the Royal Society Interface 2012 Oct 7;9(75):2653-66

A systems biology analysis of long and short-term memories of osmotic stress adaptation in fungi

You T, Ingram P, Jacobsen MD, Cook E, McDonagh A, Thorne T, Lenardon MD, de Moura APS, Romano MC, Thiel M, Stumpf M, Gow NAR, Haynes K, Grebogi C, Stark J, Brown AJ.

BMC Research Notes 2012, 5:258.

Combinatorial stresses kill pathogenic Candida species.

Kaloriti D, Tillmann A, Cook E, Jacobsen M, You T, Lenardon M, Ames L, Barahona M, Chandrasekaran K, Coghill G, Goodman D, Gow NA, Grebogi C, Ho HL, Ingram P, McDonagh A, Moura AP, Pang W, Puttnam M, Radmaneshfar E, Romano MC, Silk D, Stark J, Stumpf M, Thiel M, Thorne T, Usher J, Yin Z, Haynes K, Brown AJ.

Med Mycol. 2012 Oct;50(7):699-709

Calibrating spatio-temporal models of leukocyte dynamics against in vivo live-imaging data using approximate Bayesian computation.

Liepe J, Taylor H, Barnes CP, Huvet M, Bugeon L, Thorne T, Lamb JR, Dallman MJ, Stumpf MP.

Integr Biol (Camb). 2012 Mar;4(3):335-45

Prediction of putative protein interactions through evolutionary analysis of osmotic stress response in the model yeast Saccharomyces cerevisae.

Thorne TW, Ho HL, Huvet M, Haynes K, Stumpf MP.

Fungal Genet Biol. 2010 Dec 27.

The evolution of the Phage shock protein (Psp) response system: interplay between protein function, genomic organization and system function.

Huvet M, Toni T, Sheng X, Thorne T, Jovanovic G, Engl C, Buck M, Pinney JW, Stumpf MP.

Mol Biol Evol. 2010 Nov 8.

Estimating the size of the human interactome.

Stumpf MP, Thorne T, de Silva E, Stewart R, An HJ, Lappe M, Wiuf C.