Predictive Analytics in Genetic Engineering as an Optimization Problem
In genetic engineering, developing a breed with a desired trait is a search and optimization problem that sometimes requires many generations of field and laboratory experiments for an optimal solution to be found. The nature of the problem requires that a stochastic optimization algorithm be applie...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
INTI International University
2024
|
| Subjects: | |
| Online Access: | http://eprints.intimal.edu.my/1970/1/512 http://eprints.intimal.edu.my/1970/ http://ipublishing.intimal.edu.my/jods.html |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In genetic engineering, developing a breed with a desired trait is a search and optimization problem that sometimes requires many generations of field and laboratory experiments for an optimal solution to be found. The nature of the problem requires that a stochastic optimization algorithm be applied in the metaheuristic search rather than using a deterministic or mathematical approach. In the search for drought-tolerant cowpea, this study applied a genetic algorithm as a predictive analytics tool in the genetic engineering of three native cowpea landraces (Dan muzakkari, Gidigiwa, and Dan mesera) selected from Northern Nigeria (specifically from Kontagora in Niger State of Nigeria). The three cowpea species were subjected to mutagenic treatments using gamma irradiation and Ethyl Methane Sulphonate (EMS). Doses applied include 200, 400, 600, and 800 Gray of gamma irradiation and 0.372% v/v of EMS. Both treated and untreated cowpea landraces were planted and observed. Mutation-induced breeding aims to deepen the drought-tolerant trait of the cowpea mutants to survive conditions in drought-prone Northern Nigeria. The statistical analysis of the agro-morphological and yield parameters of the first mutant generation (M1 generation) indicates that mutagenic treatments have a positive impact on both the yield and the survival of the three landraces as all the treated landraces yielded better than the control, particularly the treatments combination of 600gray and 372% v/v of EMS. Also, the predictive outcomes of the computational simulation that was implemented in Python programming indicate that these local cultivars are developing drought-tolerant genetic variability. For the three computational experiments, the stochastic optimizer (genetic algorithm) converged at the 9412th, 9717th, and 14338th generations respectively. Such predictive analytics information is useful for guiding decision-making by researchers and breeders in the crop improvement program. |
|---|