Understanding swarming and collective motion in bacterial populations
Prof Manoj Varma (Associated Professor, Robert Bosch Centre for Cyber-Physical Systems and Centre for Nano Science and Engineering)
One of the hallmarks of a cyber-physical system is the emergence of collective intelligence from the individual elements of the system. For instance, the synthesis of a decision based on multimodal sensory information from the individual nodes of a sensor network. Complex living organisms can also be considered as cyber-physical systems because their survival depends on taking the right decisions based on the collective information gathered from their sensing organs as well as possibly from their neighbours.
In this context, bacteria and similar microbes present an excellent test-bed to study how collective intelligence emerges in a biological system. Given the fact that such systems have evolved over millions of years, the methods they use for decision making and for dealing with noise may be near-optimal and may give us valuable insights in the design of non-living cyber-physical systems.
Therefore, we would like to investigate the collective motion, referred to as swarming, of a population of bacteria, Pseudomonas Aeruginosa (PA14). Each bacteria is of the order of a micron in lengh. However, a collection of this bacteria will cover a distance of about 3 cm in a span of about 10 hours in a characteristic branching pattern. These swarming patterns display fairly robust statistical distributions of branch lengths, branch angle and so on and can also sense the presence of other bacterial colonies as far away as 1 cm. An open question in this system is how the motions of individual bacterium at such a short length scale (microns) lead to the robust collective branching pattern at a fairly large length scale (cm). There are related problems such as the extent to which a single bacterium and the swarm can sense their environment, the effect of geometry and other species and so on.
In collaboration with the laboratory of Dr. Varsha Singh at the Department of Molecular Reproduction, Development and Genetics, IISc, we will observe the swarming of PA14 on agar growth media. Current experimental data lacks the following information: (1) Information about kinetics of swarming and (2) behaviour at the single bacterium level in a swarming population. Information about these aspects will provide insights on how to model the swarming phenomena at different time and length scales. In order to make these observations, we will setup a transparent incubator (chamber with controlled temperature, humidity, O2 and CO2) at the Robert Bosch Centre. The incubator will be equipped with an imaging setup to obtain time-lapse images of the swarming pattern. We will use a combination of fluorescence and interference contrast imaging to observe the behaviour of individual bacteria. Analysis of this data will be based on in-house image/video analysis software.
In addition to these experiments, another unexplored aspect of PA14 swarming is the role of geometry, i.e. the role of shape, asymmetry or heterogeneity of the swarming field. Such studies help in understanding the extent to which a PA14 swarm can gather information about its environment. These studies will be done by fabricating swarming arenas with desired shape and heterogeneity (geometric or material). Furthermore, the large body of existing studies of PA14 swarming have been qualitative. There is virtually no data on parameters such as the minimum concentration of sensory cues which a single PA14 bacterium can sense, the spatial scale which is within the “awareness” of a single bacterium and the swarm and so on. We will address this lack of quantitative data using a combination of the tools we will develop in this project. Thus, we feel that the data we will collect during this research will fill significant gaps in the study of PA14 swarming.
A major part of this research program is the modelling of swarming, in particular the creation of an in-silico model which will reproduce the PA14 swarming patterns and their statistics accurately as well as associated behaviours such as repulsion between different colonies. In order to do this, we will employ an agent based model. The main goal of this part of the project will be to arrive at the minimal set of rules which robustly reproduces the swarming pattern accurately. The predictions of this model will be tested using the experimental tools developed in this project.