Applying Agent-Based Modeling to Studying Emergent Behaviors of the Immune System Cells

Huge amount of medical data has been generated in practical experiments which makes data analysis a challenging problem. This requires novel techniques to be developed. The improvements in computational power suggest to use computerbased modeling approaches to process a large set of data. One of the important systems in the human body to be investigated is the immune system. The previous studies of medical scientists and ongoing experiments at Karolinska Institute provide information about the human immune system. This information includes attributes of human immune system’s blood cells and the interactions between these cells. This interactions are provided as ‘if-then’ logical rules. Each rule verifies a condition on the attribute of one cell and it may initiate interaction processes to modify the attributes of other cells. A specific temporal value is associated to each process to quantify the speed of that process in the body (i.e., slow, medium, fast). We propose an agent-based model (ABM) to study human immune system cells and their interactions. The ABM is selected to overcome the complexity of large amount of data and find emergent properties and behavior patterns of the cells. Immune system cells are modeled as autonomous agents which have interactions with each other. Different values of a cell attributes define possible states of the cell and the collection of states of all cells constructs the state of the whole agent-based model. In order to consider the state transitions of the cells, we used a finite state machine (FSM). The first state is constructed from the input initial values for the cells and considering the logical time of 1. In each step, the program goes one time unit further and computes next state by applying the changes based on the cells’ interactions rules. This evolution of states in time is similar to game of life (GOL) automaton. The final model based on three modeling approaches of ABM, FSM and GOL are used to test medical hypothesis related to human immune system. This model provides a useful framework for medical scientists to do experiments on the cells’ attributes and their interaction rules. Considering a set of cells and their interactions, the proposed framework shows emergent properties and behavior patterns of the human immune system