Affiliation:
1. Department of Plant Genetics and Production, Vali-e-Asr University of Rafsanjan, Rafsanjan 7718897111, Iran
2. Department of Agronomy and Plant Breeding, Shahrood University of Technology, Shahrood P.O. Box 316-36155, Iran
3. Department of Biosystem Engineering, Faculty of Agriculture, Ferdowsi University of Mashhad, Mashhad 9177948974, Iran
4. USDA Forest Service, Northern Research Station, Hardwood Tree Improvement and Regeneration Center (HTIRC), PFEN226C, Department of Forestry and Natural Resources, Purdue University, 715 West State Street, West Lafayette, IN 47907, USA
Abstract
Linseed (Linum usitatissimum L.), a globally cultivated oilseed crop in high demand, is the focal point of our efforts aimed at improving yield production. The achievement of robust yield outcomes relies on the intricate interplay of various agronomic traits. This study, conducted over two years at a research farm in Iran, presents a comprehensive analysis evaluating diverse agronomic characteristics inherent to different linseed cultivars and hybrids. Essential parameters, including days to emergence, days to flowering, plant height, number of branches, number of capsules per plant, number of seeds per capsule, 1000-seed weight, and seed yield per plant, were examined. For predictive insights into seed yield, machine learning techniques, specifically multilayer perceptron (MLP) and multiple linear regression (MLR), were employed. The analysis of contribution percentages for each agronomic variable to linseed seed yield revealed that the number of capsules per plant emerged as the most influential factor, contributing 30.7% among the considered variables. The results indicated the superiority of MLP over MLR, with RMSE and MAPE values equaling 0.062 g/plant and 3.585%, respectively. Additionally, R2 values for training, validation, and test phases exceeded 0.97. Consequently, MLP served as a merit function in the genetic algorithm (GA), targeting the identification of optimal trait levels to maximize linseed yield. The optimization outcomes demonstrated the potential achievement of a yield of 4.40 g/plant. To attain this performance, a set of agronomic characteristic values was proposed by GA, initiating a discussion on genetic modification possibilities. The findings of this study highlight the remarkable efficacy of machine learning tools, particularly neural networks, when paired with evolutionary optimization techniques such as genetic algorithms. These methodologies prove to be invaluable assets in aiding biotechnologists as they strive to enhance the genetic makeup of products for various applications, providing unwavering reliability and invaluable guidance in the pursuit of genetic modification endeavors.
Subject
Plant Science,Agronomy and Crop Science,Food Science
Reference52 articles.
1. Genetic conversion of linseed oil from industrial to edible quality;Green;J. Am. Oil Chem. Soc.,1986
2. Muir, A.D., and Westcott, N.D. (2003). Flax: The Genus Linum, CRC Press.
3. An EMS-induced low-linolenic-acid mutant in McGregor flax (Linum usitatissimum L.);Rowland;Can. J. Plant Sci.,1991
4. Correlation and path coefficient analysis among seed yield traits and oil content in Ethiopian linseed germplasm;Tadesse;Int. J. Sustain. Crop Prod.,2009
5. Soto-Cerda, B. (2013). Association Mapping Analysis of a Core Collection of Flax (Linum usitatissimum L.), University of Manitoba.