Over the past years, numerous reports have surfaced demonstrating the outstanding superiority of combinatorial therapies over single drug treatments, one such example was the successful treatment of the human immunodeficiency virus with a combination therapy. The main problem faced when designing a multi-drug therapy is that combining a set of drugs at different possible concentrations yields a large testing parametric space, and thus the search of an optimal combination becomes a major challenge. To solve this issue, the Feedback System Control (FSC) optimization scheme has emerged as a better alternative for achieving a therapeutic goal when compared to the typical trial and error methods; FSC's primary advantage is its ability to circumvent the need for detailed information of the cellular functions of the system of interest. It has been demonstrated that only tens of iterations out of a large number of possible combinations are needed to achieve a desired response, as opposed to testing the entire search space. This effort-saving approach actively manipulates the complex biological systems as a whole, rather than controlling the system's individual intrinsic pathways.

To further exploit the capabilities of this platform, FSC has now taken advantage of the benefits offered by multivariable experimental designs such as orthogonal array composite designs; these designs are intended to draw valid correlation conclusions from an experimental data set while further minimizing the number of tests performed. In the context of FSC, they provide the initial conditions to be tested, which facilitate the development of quadratic models describing the relationship between the drug combinations and their efficacies with a reliable statistic correlation. This method is known as FSC.II.

In this project, the FSC.II methodology was used to find a drug combination for tuberculosis treatment. In six iterations, several three and four drug combinations were found to be superior to the standard regimen, which represented a drastic decrease in the number of experiments needed to find the optimal combinations for inhibiting tuberculosis infection on cell based assays. The results obtained were then verified through a colony forming unit cell based assay to verify tuberculosis killing.

These results will provide a basis of drug combinations to be tested on an animal model, where only a small number of subjects will be needed to find the optimal drug combination. Furthermore, future efforts will focus on using the FSC scheme to model the drug combination efficacy as a temporal function of a drug combination, which would allow the optimization of a drug combination efficacy over time on a single individual subject; this method would be suitable for both animal and human clinical tests and will an outstanding step towards personalized medicine.