Bayesian modelling for developmental systems biology

Lead Research Organisation: University of Warwick
Department Name: Statistics


Organ formation is controlled by the action of gene regulatory networks (GRNs) deployed in cells that are interacting with their neighbours through cell contacts and cell signalling networks. This proposal is designed to develop novel nonparametric Bayesian methods to extract the quantitative input/output relationships among gene products in the functioning GRNs that coordinate the patterning, growth and morphogenesis driving organogenesis. The Fraser Laboratory is refining and deploying technologies which will enable the reliable creation of gene expression reporters through multiplex imaging and genome profiling of discrete cell populations, creating "test points" that can be used to watch aspects of a GRN as it functions. Tagging multiple components of the GRN with distinct labels makes it feasible to read out the quantitative state of a GRN in a cell-specific and time-resolved fashion. To read out these "test points", Fraser's group is refining the equipment needed for quantitative imaging of multiple gene products over the contiguous space of a developing embryo. The novel Bayesian-based computational tools to be developed in this project will be used to reverse engineer and analyse GRNs at various scales of granularity, based on the quantitative imaging of these "test points". Ultimately, this research will permit elaboration of a more complete GRN and its linkage to key morphogenetic events by combining real-time, multiple-gene reporters, multiplex imaging and Bayesian modelling approaches. Once validated, the kit of imaging and computational tools will be broadly applicable for defining the GRN in less well-studied and accessible systems.

Planned Impact

The types of impact that we expect from this project are (using the EPSRC impact categories):

1. Impact on Knowledge. At the core of this proposal is the development of novel algorithmic tools for Bayesian computation. The contributions in this area will offer the prospect of new modalities for addressing important and challenging problems in systems biology. The collaborative research to be developed will also lead to a better understanding of the biological systems involved in organogenesis. This a critical high-order process in embryonic development and requires tools far more powerful than those conventionally deployed in cell biology, tissue culture and molecular biology studies.

2. Impact on Society, in particular impact on the understanding and treatment of cardiovascular birth defects, which affect 1-3% of live births. The work will ultimately lead to the development of a needed toolkit and its application to a highly relevant system, leading to a better understanding of birth defects associated with the cardiovascular system, as well as its normal development.

3. Impact on people. The immediate impact will be through technology transfer to research associates and PhD students in the Fraser Laboratory, as well in the reverse direction, by providing the applicant with exposure to new types of data. A future successful grant proposal will impact (i) RAs at Warwick through skills training, (ii) PhDs, notably through strong interaction with Warwick's CDTs (notably MathSys and OxWaSP) (iii) on other postdoctoral researchers through facilitating their involvement of in workshops and seminars (iv) developmental biologists and clinicians.


10 25 50
publication icon
Penfold CA (2019) Inferring Gene Regulatory Networks from Multiple Datasets. in Methods in molecular biology (Clifton, N.J.)

Description Gaussian process dynamical systems (GPDS) represent Bayesian nonparametric approaches to inference of nonlinear dynamical systems, and provide a principled framework for the learning of biological networks from multiple perturbed time series measurements of gene or protein expression. Such approaches are able to capture the full richness of complex ODE models, and can be scaled for inference in moderately large systems containing hundreds of genes. Related hierarchical approaches allow for inference from multiple datasets in which the underlying generative networks are assumed to have been rewired, either by context dependent changes in network structure, evolutionary processes, or synthetic manipulation. These approaches can also be used to leverage experimentally determined network structures from one species into another where the network structure is unknown.
Exploitation Route Collectively, these methods provide a comprehensive and flexible platform for inference from a diverse range of data, with applications in systems and synthetic biology, as well as spatiotemporal modelling of embryo development.
Sectors Agriculture, Food and Drink,Healthcare,Pharmaceuticals and Medical Biotechnology