Analysing biological information processing with mechanistic modular models

Speaker  Neil Dalchau

Host  Stephen Emmott

Affiliation  MSRC

Duration  00:57:13

Date recorded  31 October 2011

Cells, a fundamental unit of life on this planet, are able to process information deriving both from outside and within to make life-preserving decisions. Molecular signals are passed within and between cells and generated in response to extracellular stimuli, and feed into biochemical networks that carry out the major functions of an organism: energy storage, protection from pathogens, time-of-day determination, etc. Understanding biological information processing is fundamental for learning how to treat and prevent disease, or even for using biochemistry to perform computation. In this talk, I’ll show how I am using mechanistic models to reverse- and forward-engineer complex biochemistry, incorporating experimental observations and exploiting modularity in the system description. As case studies, I will describe my PhD work on circadian rhythms, then projects in the fields of immunology and synthetic biology that I have worked on at Microsoft Research.

The immune system is a complex set of mechanisms that seek to identify and rid foreign (pathogen-derived) proteins, requiring extensive information processing. I’ll introduce the first computational model of the MHC class I pathway, which has helped to elucidate how cells present a filtered snapshot of their internal contents to T lymphocytes (which blood cells), the first step in the immune system recognition of viral infection. Synthetic biology offers the potential to utilise the cellular environment for functions not intended by evolution alone. By inserting additional DNA, new functions/components can be conferred to cells that can be used to synthesise medicines and biofuels, or to teach us about how biochemistry facilitates information processing. I will show how to engineer complex spatio-temporal behaviours in populations of interacting bacteria, using intercellular signalling, inducible gene expression and a rigorous model-based design procedure.

©2011 Microsoft Corporation. All rights reserved.
> Analysing biological information processing with mechanistic modular models