Cross-cutting Methods

Membrane modelling

The transport of proteins and other biological material through constrained geometries is important in many applications, such as kidney dialysis and in the harvesting of proteins for tissue-engineering applications. In all such cases understanding the fundamental transport mechanisms is crucial in order for such technologies to be optimized or upscaled. A key challenge in designing appropriate harvesting or filtration devices is to maximize the capture efficiency while minimizing clogging of the separating device. The latter leads to significant reduction in efficiency and increase in operating costs. Recent experiments conducted by Pall Corporation suggest that membranes that comprise a pore geometry that varies with depth are able to improve separation efficiency. However, the underlying transport mechanisms responsible for the observed improvements are currently not understood since experiments are unable to probe the microscopic behaviour.


Scaling up Markov chain Monte Carlo methods

Markov chain Monte Carlo methods (MCMC) are the cornerstone of Bayesian inference in complex models. Massive datasets (as in genetics) and large-scale, complex models (as in ecology) are raising new issues for MCMC. We investigate ways to scale up these algorithms with growing data and dimension.

Individual-based modelling

The aim of this work is to develop novel tools and methods for representing collective behaviour. It is also provides the techniques necessary to underpin the application work in a number of our associated projects.

Visual Diagnostics for MCMC (BACK-FILZ)

Markov Chain Monte Carlo (MCMC) sampling techniques play a central role in estimating parameter values in complicated models. Advances in MCMC techniques and software have opened up a new set of tools and users can now confront models with heterogeneous data sets. Like all numerical methods, MCMC techniques approximate a target distribution (e.g. a joint posterior distribution in a Bayesian Analysis) and require a number of statistical diagnostics to assess the use of these algorithms.

Visual analytics for bioinformatics (PIGGLEDY)

Bioinformatics research yields vast amounts of data on the relationships amongst proteins. Exploring the connections between proteins (e.g. phylogenetic, structural or functional similarity) has many applications in academic sciences and industrial applications, and education. However, a suitable interface does not exist. Previous visualisations (of which there are very few considering the wealth of data) have primarily used single, static views of the data, employed standard data representations or focused on visualising the sequences.

Functional Curation

The key idea underpinning functional curation is that when mathematical and computational models are being developed, a primary goal should be the continuous comparison of those models against experimental data. When computational models are being re-used in new studies, it is similarly important to check that they behave appropriately in the new situation to which you're applying them. To achieve this goal, it's a pre-requisite to be able to replicate in-silico precisely the same protocols used in an experiment of interest. We are therefore developing a language for describing rich 'virtual experiment' protocols, and software for running these on compatible models.

Syndicate content