Modeling the Bending Strength of MDF Faced, Polyurethane Foam-Cored Sandwich Panels Using Response Surface Methodology (RSM) and Artificial Neural Network (ANN)

Optimization of preparation conditions for Salsola laricifolia protoplasts using response surface methodology and artificial neural network modeling

BACKGROUND

Salsola laricifolia is a typical C3-C4 typical desert plant, belonging to the family Amaranthaceae. An efficient single-cell system is crucial to study the gene function of this plant. In this study, we optimized the experimental conditions by using Box-Behnken experimental design and Response Surface Methodology (RSM)-Artificial Neural Network (ANN) model based on the previous studies.

RESULTS

Among the 17 experiment groups designed by Box-Behnken experimental design, the maximum yield (1.566 × 106/100 mg) and the maximum number of viable cells (1.367 × 106/100 mg) were obtained in group 12, and the maximum viability (90.81%) was obtained in group 5. Based on these results, both the RSM and ANN models were employed for evaluating the impact of experimental factors. By RSM model, cellulase R-10 content was the most influential factor on protoplast yield, followed by macerozyme R-10 content and mannitol concentration. For protoplast viability, the macerozyme R-10 content had the highest influence, followed by cellulase R-10 content and mannitol concentration. The RSM model performed better than the ANN model in predicting yield and viability. However, the ANN model showed significant improvement in predicting the number of viable cells. After comprehensive evaluation of the protoplast yield, the viability and number of viable cells, the optimal results was predicted by ANN yield model and tested. The amount of protoplast yield was 1.550 × 106/100 mg, with viability of 90.65% and the number of viable cells of 1.405 × 106/100 mg. The corresponding conditions were 1.98% cellulase R-10, 1.00% macerozyme R-10, and 0.50 mol L-1 mannitol. Using the obtained protoplasts, the reference genes (18SrRNA, β-actin and EF1-α) were screened for expression, and transformed with PEG-mediated pBI121-SaNADP-ME2-GFP plasmid vector. There was no significant difference in the expression of β-actin and EF1-α before and after treatment, suggesting that they can be used as internal reference genes in protoplast experiments. And SaNADP-ME2 localized in chloroplasts.

CONCLUSION

The current study validated and evaluated the effectiveness and results of RSM and ANN in optimizing the conditions for protoplast preparation using S. laricifolia as materials. These two methods can be used independently of experimental materials, making them suitable for isolating protoplasts from other plant materials. The selection of the number of viable cells as an evaluation index for protoplast experiments is based on its ability to consider both protoplast yield and viability. The findings of this study provide an efficient single-cell system for future genetic experiments in S. laricifolia and can serve as a reference method for preparing protoplasts from other materials.

An Improved Optimization Model to Predict the MOR of Glulam Prepared by UF-Oxidized Starch Adhesive: A Hybrid Artificial Neural Network-Modified Genetic Algorithm Optimization Approach

The purpose of the present article is to study the bending strength of glulam prepared by plane tree (Platanus Orientalis-L) wood layers adhered by UF resin with different formaldehyde to urea molar ratios containing the modified starch adhesive with different NaOCl concentrations. Artificial neural network (ANN) as a modern tool was used to predict this response, too. The multilayer perceptron (MLP) models were used to predict the modulus of rapture (MOR) and the statistics, including the determination coefficient (R²), root mean square error (RMSE), and mean absolute percentage error (MAPE) were used to validate the prediction. Combining the ANN and the genetic algorithm by using the multiple objective and nonlinear constraint functions, the optimum point was determined based on the experimental and estimated data, respectively. The characterization analysis, performed by FTIR and XRD, was used to describe the effect of the inputs on the output. The results indicated that the statistics obtained show excellent MOR predictions by the feed-forward neural network using Levenberg-Marquardt algorithms. The comparison of the optimal output of the actual values obtained by the genetic algorithm resulting from the multi-objective function and the optimal output of the values estimated by the nonlinear constraint function indicates a minimum difference between both functions.