Dr. Mark Read

Research Fellow, Charles Perkins Centre, The University of Sydney
Research Gate
My Research Gate

Research

My research lies on the interdisciplinary boundary between computational sciences, and immunology, specifically addressing how these disciplines can advance one another.

The following is a list of current research projects that I engage with.

“Integrative simulation techniques to disentangle host-diet-gut microbiome interactions.” 
With A/Prof Andrew Holmes (USyd), Prof Stephen Simpson (Usyd), Prof David Raubenheimer (Usyd). 
The microbiome is increasingly implicated in our health status, but the factors underlying the composition of this complex cross-feeding community remain elusive. We use novel simulation techniques to determine intervention strategies to manipulate the microbiome to our benefit.
“Determining optimal search strategies through theory, functional analysis and simulation.” 
With Dr Maté Biro (UNSW), Dr Greg Rice (Waterloo, Canada). 
We are developing technologies to determine the best way for motile agents to find their targets, by they mates, food or shelter. This has applications in immunology, ecology, zoology and in swarm robotic systems.
“Unravelling the dynamics of haematopoietic stem cells.”
With Dr. Ben Roediger, Dr. Chris Jolly (Centenary Inst)
These stem cells ultimately give rise to all blood and immune system cells, their function is critical to our health. We are using simulation to reconcile seeming contradictions around their dynamics, and determining how their behaviour can lead to either homeostasis or cancer.
“Factors underlying neutrophil swarming elucidated through simulation.”
With Dr Tatyana Chtanova (Garvan), Prof Jon Timmis (University of York, UK). 
Neutrophils, key immune system cells, exhibit a striking swarming motility pattern in response to pathogens and injury. We are using simulation techniques to recreate these dynamics and thereby understand their emergence.
“Methdological advancements in simulation-based science”
With Prof Jon Timmis, Dr. Kieran Alden (both University of York, UK)
Simulation is increasingly employed in biological science, it facilitates interventions and a degree of experimental precision not possible in the real world. We are investigating machine learning and optimisation techniques to advance simulation-based science, yielding greater insights and ensuring simulations are representative of the biological systems we are investigating.
“Mapping dynamic immunity: next-generation computational approaches to track the evolution of immune responses.”
With A/Prof Irena Koprinska, A/Prof Uwe Roehm, Dr. Thomas Asshurt, Prof Nick King (all Usyd).
Cytometry technology, whereby individual cells can be characterised in up to 45 dimensions, offers a powerful means of analysing the immune response. We are developing novel machine learning techniques (clustering) that can accommodate the challenge of 45 dimensional temporal data, and map out how our immune systems respond to different challenges.
“Predicting Asthma and other respiratory diseases through machine learning.”
With A/Prof Irena Koprinska (USyd), Dr. Cindy Thamrin (Woolcock Inst). 
We are building more powerful predictors of patient disease, and when they will experience clinical exacerbations, such that they can better manage their health.
“Predicting potential weightloss from faecal samples through machine learning.”
With A/Prof Irena Koprinska (USyd), Dr. Nick Fuller (Boden Inst), A/Prof Andrew Holmes (USyd).  
People respond very differently to a given diet, with only some experiencing success. Here we use classification and regression of high-dimensional microbiome sequencing data to predict the best dietary intervention for a given patient.

 

Computational Immunology

I am an visiting member of YCIL (York Computational Immunology Lab). My research in this area has included the modelling and simulation of EAE, a mouse proxy for multiple sclerosis, and modelling the chemotactic soluble factors responsible for the phenomenon of neutrophil swarming in response to tissue damage. The EAE work has been conducted in collaboration with Dr. Vipin Kumar at the TPMIS, San Diego. The neutrophil swarming work is conducted in collaboration with Dr. Tatyana Chtanova from the Garvan Institute, Sydney.

Computational methods are emerging as a valuable complement to the traditional wet-lab approaches to exploring immunology. They permit the consolidation of a wide variety of data, potentially originating from different labs, models and investigative methods, and provide an overview of the system of interest. They permit the formulation of novel hypotheses of how the immune system operates, and their evaluation in the context of established knowledge.

My research is primarily concerned with understanding the link between simulation and the real world, establishing confidence that the results of simulation are representative of their real-world systems. My research involves the development and application of statistical and modelling methods in the creation and analysis of simulations to meet this goal.

 

Nature-Inspired Computation

An exciting strand of computer science and engineering entails addressing engineering problems (such as balancing conflicting requirements whilst finding near-optimal solutions, or engineering complex behaviours through integration of simple behaviours across a large population of actors) though inspiration adopted from nature. Natural systems often represent highly robust solutions to complex problems. Take the immune system, which can direct its destructive potential towards harmful organisms whilst maintaining the health of the host. It does this in a completely de-centralised manner, there is no single variety of cell that commands the actions of others.

My research entails identifying qualities in natural systems that can serve as inspiration for solutions to complex engineering problems. I am particularly interested in how such solutions are derived from their natural inspiration; natural systems are rarely simple to comprehend, and oversimplification can lead to a failure to capture attractive properties.

 

Swarm Robotics

Swarm robots marries nature-inspired computation with a real-world problem domain. The problem is the control of group level robotic behaviour through the manipulation of (relatively) simple single-robot function. Swarms of robots are suitable for addressing problems where the danger of losing a single robot is high, or where a present in a wide physical area is required. Interesting research is being conducted within CoCoRo on how group-level decision making can emerge from perceptions and interactions of single group members.