Forecasting and optimizing Agrobacterium-mediated genetic transformation via ensemble model- fruit fly optimization algorithm: A data mining approach using chrysanthemum databases | JoVE Visualize

Area of Science:

Plant Biotechnology
Molecular Biology
Computational Biology

Background:

Optimizing gene transformation protocols is essential for genetic engineering and genome editing but is often complex, costly, and time-consuming.
Novel computational approaches, specifically machine learning, are needed to analyze gene transformation data and improve efficiency.

Purpose of the Study:

To develop and evaluate machine learning models for forecasting Agrobacterium-mediated gene transformation in chrysanthemum.
To create a hybrid model by fusing individual machine learning models and optimizing it using a metaheuristic algorithm.

Main Methods:

Developed three individual machine learning models: Multi-Layer Perceptron (MLP), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Radial Basis Function (RBF).
Input variables included Agrobacterium strain, optical density (OD), co-culture period (CCP), and various antibiotics (kanamycin, vancomycin, cefotaxime, hygromycin, carbenicillin, geneticin, ticarcillin, paromomycin).
Fused the best-performing individual models using a bagging method and optimized the ensemble model with the Fruit Fly Optimization Algorithm (FOA).

Main Results:

The ensemble model demonstrated superior performance with an R2 of 0.83, outperforming individual models (MLP: 0.63, RBF: 0.69, ANFIS: 0.74) on the validation set.
The optimized hybrid model predicted a maximum gene transformation efficiency of 37.54% under specific conditions (EHA105 strain, 0.9 OD600, 3.8 days CCP, and defined concentrations of selection antibiotics).
Sensitivity analysis revealed the order of importance for input variables: Agrobacterium strain > CCP > K > CF > VA > P > OD > CA > H > TI > G.

Conclusions:

The developed hybrid ensemble model-FOA is an accurate and reliable computational tool for optimizing Agrobacterium-mediated gene transformation.
This approach can significantly aid future genetic engineering and genome editing studies by streamlining protocol development and improving success rates.