Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learning
Developing new tissue culture protocols and optimization of plant media composition for particular plant species are precluded by the higher labor, time and cost. The usage of machine learning becomes rewarding tools for efficient, cost-effective method of optimization of growth media composition fo...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2020-12-01
|
| Series: | Information Processing in Agriculture |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2214317319300940 |
| _version_ | 1797726838365093888 |
|---|---|
| author | Sachithri P. Munasinghe Seneviratnege Somaratne Shyama R. Weerakoon Chandani Ranasinghe |
| author_facet | Sachithri P. Munasinghe Seneviratnege Somaratne Shyama R. Weerakoon Chandani Ranasinghe |
| author_sort | Sachithri P. Munasinghe |
| collection | DOAJ |
| description | Developing new tissue culture protocols and optimization of plant media composition for particular plant species are precluded by the higher labor, time and cost. The usage of machine learning becomes rewarding tools for efficient, cost-effective method of optimization of growth media composition for desires growth performance. Difficulties are often encountered in determination of media composition for raising calli from explants of woody Gyrinops walla. This paper provides the finding of the study performed on modeling and optimization of chemical composition of culture media for raising callus from explants of G. walla. Number of experiments were conducted to raise calli from sterilized leaf and node explants with varying concentrations of 1- Naphthaleneacetic acid (NAA), 6-Benzylaminopurine (BAP) and coconut water in Murashige and Skoog medium (MSM) and Woody plant medium (WPM) were conducted. Artificial Neural Networks (ANNs) were developed to predict the best media composition on callus weight (CW), callus induction percentage (CIP) and days taken to initiate callus (DC). Callus formation was observed in all the experiments except for controls. Three ANN models were developed to predict the best media composition for CW, CIP and DC. The results of ANN models predict optimum media composition during the training, validation and testing with satisfactory levels of performance. The R2 values of the best ANN models; CWnet, CIPnet and DCnet were 0.95, 0.95 and 0.88 respectively. The mean sums of square errors (RMSRs) of CWnet, CIPnet and DCnet were 0.1410, 0.1391 and 0.1961 and mean residual (MRs) were 0.0199, 0.0194 and 0.0385. The corrected Akaike information criterion (AICc) of the models was −331.431, −254.926 and −210.777, and Schwarz Bayesian information criterion (SBC) were −490.495, −454.229 and −382.123. The predicted best media consisting MSM + 3.0 mg/l NAA + 0.5 mg/l BAP and MSM + 1.5 mg/l NAA + 0.5 mg/l BAP were the optimum combination for leaves and nodal explants respectively. The AICc and SBC of CWnet indicated better performance over the rest of the models in predicting media composition. The developed ANN models indicated a novel approach using ANN models in predicting media composition for raising callus from woody G. walla by hypothetical chemical composition saving cost, time and labor. Findings suggest that application of artificial intelligence in predicting media composition for micro-propagation of woody plants in future studies. |
| first_indexed | 2024-03-12T10:51:15Z |
| format | Article |
| id | doaj.art-2064b097b719412d85546714115863e9 |
| institution | Directory Open Access Journal |
| issn | 2214-3173 |
| language | English |
| last_indexed | 2024-03-12T10:51:15Z |
| publishDate | 2020-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Information Processing in Agriculture |
| spelling | doaj.art-2064b097b719412d85546714115863e92023-09-02T06:56:44ZengElsevierInformation Processing in Agriculture2214-31732020-12-0174511522Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learningSachithri P. Munasinghe0Seneviratnege Somaratne1Shyama R. Weerakoon2Chandani Ranasinghe3Department of Botany, The Open University of Sri Lanka, P.O. Box 21, Nawala, Nugegoda 10250, Sri LankaDepartment of Botany, The Open University of Sri Lanka, P.O. Box 21, Nawala, Nugegoda 10250, Sri LankaDepartment of Botany, The Open University of Sri Lanka, P.O. Box 21, Nawala, Nugegoda 10250, Sri Lanka; Corresponding author.Department of Chemistry, The Open University of Sri Lanka, P.O. Box 21, Nawala, Nugegoda 10250, Sri LankaDeveloping new tissue culture protocols and optimization of plant media composition for particular plant species are precluded by the higher labor, time and cost. The usage of machine learning becomes rewarding tools for efficient, cost-effective method of optimization of growth media composition for desires growth performance. Difficulties are often encountered in determination of media composition for raising calli from explants of woody Gyrinops walla. This paper provides the finding of the study performed on modeling and optimization of chemical composition of culture media for raising callus from explants of G. walla. Number of experiments were conducted to raise calli from sterilized leaf and node explants with varying concentrations of 1- Naphthaleneacetic acid (NAA), 6-Benzylaminopurine (BAP) and coconut water in Murashige and Skoog medium (MSM) and Woody plant medium (WPM) were conducted. Artificial Neural Networks (ANNs) were developed to predict the best media composition on callus weight (CW), callus induction percentage (CIP) and days taken to initiate callus (DC). Callus formation was observed in all the experiments except for controls. Three ANN models were developed to predict the best media composition for CW, CIP and DC. The results of ANN models predict optimum media composition during the training, validation and testing with satisfactory levels of performance. The R2 values of the best ANN models; CWnet, CIPnet and DCnet were 0.95, 0.95 and 0.88 respectively. The mean sums of square errors (RMSRs) of CWnet, CIPnet and DCnet were 0.1410, 0.1391 and 0.1961 and mean residual (MRs) were 0.0199, 0.0194 and 0.0385. The corrected Akaike information criterion (AICc) of the models was −331.431, −254.926 and −210.777, and Schwarz Bayesian information criterion (SBC) were −490.495, −454.229 and −382.123. The predicted best media consisting MSM + 3.0 mg/l NAA + 0.5 mg/l BAP and MSM + 1.5 mg/l NAA + 0.5 mg/l BAP were the optimum combination for leaves and nodal explants respectively. The AICc and SBC of CWnet indicated better performance over the rest of the models in predicting media composition. The developed ANN models indicated a novel approach using ANN models in predicting media composition for raising callus from woody G. walla by hypothetical chemical composition saving cost, time and labor. Findings suggest that application of artificial intelligence in predicting media composition for micro-propagation of woody plants in future studies.http://www.sciencedirect.com/science/article/pii/S2214317319300940Gyrinops walla GaetnerTissue cultureCoconut waterWoody plantArtificial neural network |
| spellingShingle | Sachithri P. Munasinghe Seneviratnege Somaratne Shyama R. Weerakoon Chandani Ranasinghe Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learning Information Processing in Agriculture Gyrinops walla Gaetner Tissue culture Coconut water Woody plant Artificial neural network |
| title | Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learning |
| title_full | Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learning |
| title_fullStr | Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learning |
| title_full_unstemmed | Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learning |
| title_short | Prediction of chemical composition for callus production in Gyrinops walla Gaetner through machine learning |
| title_sort | prediction of chemical composition for callus production in gyrinops walla gaetner through machine learning |
| topic | Gyrinops walla Gaetner Tissue culture Coconut water Woody plant Artificial neural network |
| url | http://www.sciencedirect.com/science/article/pii/S2214317319300940 |
| work_keys_str_mv | AT sachithripmunasinghe predictionofchemicalcompositionforcallusproductioningyrinopswallagaetnerthroughmachinelearning AT seneviratnegesomaratne predictionofchemicalcompositionforcallusproductioningyrinopswallagaetnerthroughmachinelearning AT shyamarweerakoon predictionofchemicalcompositionforcallusproductioningyrinopswallagaetnerthroughmachinelearning AT chandaniranasinghe predictionofchemicalcompositionforcallusproductioningyrinopswallagaetnerthroughmachinelearning |