Artificial Neural Networks QSAR Rhodiola Rosea In Vitro
Issue Date:
2018
Publisher:
Institute of Mathematics and Informatics Bulgarian Academy of Sciences
Citation:
Serdica Journal of Computing, Vol. 12, No 3, (2018), 191p-218p
Abstract:
The aim is to find a cost-effective method of identifying nutrient media producing identical plants with maximum performance in terms of the bioactive substances contained in them. An adaptation of QSAR is used. The “spatial structure of the chemical component” is replaced with “the multidimensional structure of the nutrient medium” or with the treated day schemes. For each process, a separate forecast is made. All nutrition media produced in silico are based on the ranges of phytonutrient hormones in biotechnological experiments. We found 43 theoretical combinations of media with more than 80% success under conditions of limited resources in the price range of [0–1,5] euro/liter. The obtained results can be used as: a theoretical guideline for determining the optimal nutrient media and combinations; to the study of other medicinal plants in order to establish effective biotechnological schemes for growth and rooting that are also cost-effective; using ANN, taking into account the species and the ecotype.
