The compound combination optimization only utilized prospectively obtained morphological data of 59 red spinach treated according to the OACD combinations, and did not require any pre-existing datasets for the optimization workflow. With these data, WisDM Green rapidly correlated the relationship between the compound combinations and the %Yield of each treated red spinach via a second order quadratic series to predict the %Yield of all 6,561 possible combinations at three concentration levels. WisDM Green utilized a resolution IV 59-combination OACD, which requires a small but representative sample size, to pinpoint unforeseen compound interactions. Aside from the chosen design, alternative higher resolution OACD’s may enhance the predictability of WisDM Green by screening more combinations, which may lead to higher costs. Thus, a balance between the efficient use of resources and predictability must be carefully considered \cite{Lim_2019}. Importantly, WisDM Green pinpointed unforeseen compound interactions and concentration ratios that positively impacted %Yield via 2-compound combinations. Therefore, this strategy represents a first step towards improving peat moss formulation by optimizing compound combinations and their concentrations to positively impact %Yield while simultaneously mitigating wastage.
Given the agnostic nature of WisDM Green implementation, it can potentially be expanded to other applications in food production and wider farming communities. These may include cell culture media optimization for cell-based meats, beverage compound selection, viticulture, space farming, and other applications \cite{Massa_2017}. WisDM Green is also able to prioritize compound combinations that do or do not contain certain agents. Examples include selecting an optimal combination that does not contain animal products, or perhaps contains only vegan diet-compliant compounds, among other criteria. Moreover, WisDM Green allows multi-parametric optimization to determine the most suitable combinations for a specific desired outcome. For instance, this approach may be used to pinpoint compound combinations that optimize for biological yield without compromising the nutritional content. WisDM Green may also be applicable towards improving the yield of plant-derived compounds for cosmetic or drug synthesis, for example. Furthermore, WisDM Green can be rapidly re-implemented to account for evolving factors such as reagent availability, cost, effectiveness, climate and environmental change, user requirement, and many other parameters (Figure 2).
Concentration-dependent Synergy and Sustainable Farming
In this proof-of-concept study, we harnessed a platform technology that we have previously applied in drug combination optimization towards plant biological yield optimization. Similarly, we also explored approaches used in drug development, such as the Bliss independence model, to assess synergies observed in WisDM Green-pinpointed compound combinations \cite{Liu_2018,Poon_2021}. In the context of sustainability, rationally optimizing yield enhancement may be closely interconnected with concentration-dependent synergy. Appropriately adjusting concentration ratios to achieve optimal outcomes may substantially reduce the use of compounds, and also further reduce the reliance on fertilizer-driven approach to increase yield. For example, the excessive use of fertilizer has affected aquatic life and increased greenhouse gas emissions \cite{Khan_2018,Bijay_Singh_1995,Qadri_2019,Huang_2017,Malyan_2019,Sedlacek_2020,_gmundarson_2020}. Though efforts have been made to protect the environment via approaches, such as controlled-release fertilizer, properly adjusting the concentration ratios of compounds may result in improved outcomes and reduction in fertilizer usage, which may lead to leaching \cite{Li_2018,Sikora_2020,Xiao_2019,_gmundarson_2020,Puga_2020}.