Tumours behave as moving targets that can evade chemotherapeutic treatments by rapidly acquiring resistance via various mechanisms. In Balaz et al. (2021, Biosystems; 199:104290) we initiated the development of the agent-based open-ended evolutionary simulator of novel drug delivery systems (DDS). It is an agent-based simulator where evolvable agents can change their perception of the environment and thus adapt to tumour mutations. Here we mapped the parameters of evolvable agent properties to the realistic biochemical boundaries and test their efficacy by simulating their behaviour at the cell scale using the stochastic simulator, STEPS. We show that the shape of the parameter space evolved in our simulator is comparable to those obtained by the rational design.

The abilities of organisms to discern, categorize and act on external cues are very sophisticated and are based on a number of underlying processes. To investigate the development of these abilities, we designed a new evolutionary agentbased system where agents start with no executable functions nor with the inherent ability to recognize other elements in their environment. Instead, the agents can only perceive externally visible properties (phenotype) of other agents initially. Over the course of evolution they gradually construct their own reactions to perceived properties. To minimize constraints on the building of adaptations, the only measure of success in our model is the agents’ ability to survive and reproduce. To survive, agents need to collect energy by learning to recognize and feed on either other agents or on “primary food”. We ran a series of experimental runs where we compared evolutionary development of agents between two settings: s + where agents are allowed to develop awareness of their environment and s − where they are completely ignorant of it. In the case of s + evolution, the system settled on cyclic interdependent swings between a large number of prey and much smaller number of predators, as in real ecological systems. Our results show that in s + setting, the agent’s early evolutionary focus is to, as soon as possible, expand their ability to perceptively assimilate environment and to functionally categorize environment by developing a variety of adaptive responses to newly assimilated environmental properties.