Myxobacteria are fascinating creatures: when food is plentiful, they feed as a multicellular swarm. Though each bacterium is autonomous with respect to metabolism and reproduction, together they make up a multicellular organism. A swarm is a predatory collective that moves and feeds cooperatively, hunting together and pooling extracellular enzymes when digesting prey bacteria. When food runs short, the hundreds of thousands of swarm cells change their behavior to initiate a self-organized program that builds densely packed aggregates, called fruiting bodies, within which rod shaped cells differentiate into spherical, starvation-resistant spores. If a moving object, such as the leg of an insect, comes in contact with a fruiting body, the entire package of spores -- the fruiting body -- will likely be picked up and carried as a unit by the insect. This way, if carried to a new food source (toward which the insect was heading), the thousands of spores can germinate and emerge as an "instant" swarm, rather than having to re-establish a swarm from a single cell. Understanding the self-organization of swarms and fruiting bodies has potential impact for understanding the development of multi-cellular organisms, possibly including certain birth defects.
We will develop a multiscale 3 dimensional (ED) computational model of Myxobacterial fruiting body formation based on very short range (cell contact) interactions, differentiation and motility. We will analyze the effect of particular mutations on fruiting body development, comparing experiments with simulations for improving the model. Single scale models of biochemical and cellular networks are unable to capture the complexity, for even very basic biological phenomena occurr over diverse space and time scales. This is why it is crucial to determine how best to combine models at different scales and how the combination of several different types of models impacts the accuracy of the general multiscale model. These models are typically run as an iterative workflow over a distributed system. Management of these distributed workflows can be complicated to an end user and necessitates a convenient descriptive model that is data and result oriented, rather than task oriented. We will implement our multiscale models in a distributed problem solving environment. This environment will allow easy configuration and manipulation of the workflows to perform analysis of biochemical and cellular networks through an extension to the Systems Biology Markup Language (SBML). Proposed 3D computational model will also serve as a tool for analyzing mechanisms for building other multicellular structures dependent on cell contact signaling.
We will disseminate all results in an easily deployable bundle of the proposed Systems Biology Toolkit (SBT) to model molecular and subcellular levels, and the CellAggregate package to model the multicellular level. The interdisciplinary research team encompasses a computer scientist, a mathematician, a biophysicist, and a developmental biologist and biochemist, in a multi-institution collaboration including Notre Dame, Stanford, and Los Alamos National Lab.
