Pseudomonas aeruginosa is an organism notable for its ubiquity in the ecosystem and its resistance to antibiotics. It is an environmental bacterium that is a common cause of hospital-acquired infections. Identifying its survival mechanism is critical for designing preventative and curative measures. Also, understanding this mechanism is beneficial because P. aeruginosa and other related organisms are capable of bioremediation. To address this practical problem, we proceeded by decomposition into multiple learnable components, two of which are presented in this paper. With unlabeled data collected from P. aeruginosa gene expression response to low nutrient water, a Bayesian Machine Learning methodology was implemented, and we created an optimal regulatory network model of the survival mechanism. Subsequently, node influence techniques were used to computationally infer a group of twelve genes as key orchestrators of the observed survival phenotype. These results are biologically plausible, and are of great contribution to the overall goal of apprehending P. aeruginosa survival mechanism in nutrient depleted water environment.