Title |
Development of Prediction Model and Experimental Validation in Predicting the Curcumin Content of Turmeric (Curcuma longa L.)
|
---|---|
Published in |
Frontiers in Plant Science, October 2016
|
DOI | 10.3389/fpls.2016.01507 |
Pubmed ID | |
Authors |
Abdul Akbar, Ananya Kuanar, Raj K. Joshi, I. S. Sandeep, Sujata Mohanty, Pradeep K. Naik, Antaryami Mishra, Sanghamitra Nayak |
Abstract |
The drug yielding potential of turmeric (Curcuma longa L.) is largely due to the presence of phyto-constituent 'curcumin.' Curcumin has been found to possess a myriad of therapeutic activities ranging from anti-inflammatory to neuroprotective. Lack of requisite high curcumin containing genotypes and variation in the curcumin content of turmeric at different agro climatic regions are the major stumbling blocks in commercial production of turmeric. Curcumin content of turmeric is greatly influenced by environmental factors. Hence, a prediction model based on artificial neural network (ANN) was developed to map genome environment interaction basing on curcumin content, soli and climatic factors from different agroclimatic regions for prediction of maximum curcumin content at various sites to facilitate the selection of suitable region for commercial cultivation of turmeric. The ANN model was developed and tested using a data set of 119 generated by collecting samples from 8 different agroclimatic regions of Odisha. The curcumin content from these samples was measured that varied from 7.2% to 0.4%. The ANN model was trained with 11 parameters of soil and climatic factors as input and curcumin content as output. The results showed that feed-forward ANN model with 8 nodes (MLFN-8) was the most suitable one with R(2) value of 0.91. Sensitivity analysis revealed that minimum relative humidity, altitude, soil nitrogen content and soil pH had greater effect on curcumin content. This ANN model has shown proven efficiency for predicting and optimizing the curcumin content at a specific site. |
X Demographics
Geographical breakdown
Country | Count | As % |
---|---|---|
Switzerland | 1 | 100% |
Demographic breakdown
Type | Count | As % |
---|---|---|
Members of the public | 1 | 100% |
Mendeley readers
Geographical breakdown
Country | Count | As % |
---|---|---|
Unknown | 57 | 100% |
Demographic breakdown
Readers by professional status | Count | As % |
---|---|---|
Student > Ph. D. Student | 10 | 18% |
Researcher | 9 | 16% |
Student > Bachelor | 8 | 14% |
Student > Master | 4 | 7% |
Student > Doctoral Student | 3 | 5% |
Other | 8 | 14% |
Unknown | 15 | 26% |
Readers by discipline | Count | As % |
---|---|---|
Agricultural and Biological Sciences | 14 | 25% |
Engineering | 9 | 16% |
Biochemistry, Genetics and Molecular Biology | 3 | 5% |
Pharmacology, Toxicology and Pharmaceutical Science | 2 | 4% |
Neuroscience | 2 | 4% |
Other | 8 | 14% |
Unknown | 19 | 33% |