The microbiome is an interconnected network of microorganisms, which exist and influence a wide array of natural and synthetic environments. Genetic information is constantly spread across the members of the microbial community in a process called horizontal gene transfer, causing exposure of genetic alterations and modifications to all members of the community. In order to accurately and effectively engineer microbiomes, genetic modifications must be introduced to certain species, as selectivity is a key factor in creating and fixing functional abilities within microbial environments. Moreover, introduction of genes into unwanted hosts may cause unprecedented ecological impacts, posing a major biosafety issue. Technologies in the field are usually experimentally developed for a specific host or environment, and the lack of automization and generalization limit them to a specific microbiome. Additionally, they only deal with the transformation process itself at best and do not modulate the different elements of the genetic material, neglecting considerations related to the interactions between the new genetic material and the population. This work presents a set of computational models that automatically design a microbiome-specific plasmid that is selectively expressed in certain parts of the bacterial population. The underlying algorithm fine-tunes genetic information to be optimally expressed in the wanted hosts of the plasmid, while simultaneously impairing expression in unwanted hosts. We take into account and selectively optimize the main elements linked to gene expression and heredity. In addition, we have provided both in-silico and in-vitro analysis supporting our claim. This study was part of the TAU IGEM 2021 project (https://2021.igem.org/Team:TAU_Israel ).