skip to content

Cambridge NERC Doctoral Training Partnerships

Graduate Research Opportunities
Flow streamlines induced by swimming Volvox alga (courtesy of Ray Goldstein, DAMTP)
Brief summary: 
Applying the tools of machine learning to understand how marine microorganisms optimise their transport in complex fluid environments
Importance of the area of research concerned: 
Microorganisms constitute the bulk of the biomass of our planet and play a critical role in the life of higher organisms by performing chemical reactions and providing nutrients. Marine microorganisms in particular, ranging from single-celled bacteria to multicellular alga, participate in the (marine) cycle of many chemical elements, including carbon and oxygen. This research project will investigate the physical mechanisms dictating the selection of swimming behaviours by marine microorganisms, in particular looking for “optimal” locomotion modes in complex fluid environments. Our results will lead to a better understanding of the spatio-temporal distribution of marine microorganisms in the oceans.
Project summary : 
The behaviour of marine microorganisms is strongly influenced by the physical constraints of their habitat. The most prominent of these constraints is the presence of a surrounding fluid. Most microorganisms are motile and this motility is achieved by the coordinated motion of flagella and cilia, slender organelles a few microns long whose time-varying actuation creates hydrodynamic stresses and leads to propulsion. In this project, the student will use reinforcement learning, one of the areas of machine learning, to develop and solve mathematical models of motile marine microorganisms that are optimising their motion in complex, time-varying environments. Our ultimate goal is to understand the physical basis for the cellular dynamics observed in nature.
What will the student do?: 
The student will develop and solve mathematical models for the dynamics of swimming cells in time-varying oceanic flows. This will involve developing in parallel (i) an understanding of the hydrodynamics of flagellated and ciliated cells in time-varying viscous flows (which govern the dynamics of cells in the ocean) and (ii) an understanding of Markov decision processes (which provide the mathematical foundation for reinforcement learning algorithms). The student will then study in silico the active dynamics of cells whose swimming behaviour are governed by different biophysical rewards, including displacement, residence time and efficiency. By comparing the results of the machine learning predictions with published experimental results, we will discover the underlying trade-offs and the optimisation of marine microorganisms transport in complex fluids. The student should have a strong background in applied mathematics, particularly fluid mechanics, and a strong interest in applications of mathematical methods to understand the natural world.
References - references should provide further reading about the project: 
J. S. Guasto, R. Rusconi, R. Stocker (2012) Fluid mechanics of planktonic microorganisms, Annual Review of Fluid Mechanics 44, 373-400
S. Colabrese, K. Gustavsson, A. Celani, and L. Biferale (2017) Flow Navigation by Smart Microswimmers via Reinforcement Learning Phys. Rev. Lett. 118, 158004
A. C H. Tsang, P. W. Tong, S. Nallan, and O. S. Pak (2020) Self-learning how to swim at low Reynolds number, Phys. Rev. Fluids 5, 074101
You can find out about applying for this project on the Department of Applied Mathematics and Theoretical Physics (DAMTP) page.
Prof Eric Lauga
Department of Applied Mathematics and Theoretical Physics PhD Admissions