Prediction of genetic value in Eucalyptus grandis x E. urophylla hybrid progeny test using the REML-BLUP method and by means of artificial neural networks: deep learning

• Grant number: 21/07776-0
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: November 01, 2021
• End date: October 31, 2022
• Field of knowledge: Agronomical Sciences - Agronomy - Crop Science
• Principal Investigator: Rafael Simões Tomaz
• Grantee: Vitória Costa Mingoranci
• Host Institution: Faculdade de Ciências Agrárias e Tecnológicas. Universidade Estadual Paulista (UNESP). Campus de Dracena. Dracena , SP, Brazil

Abstract

Eucalypt is currently the most widely planted exotic species in Brazil due to its high commercial value. With the establishment of the species, several research and breeding programs were initiated with the purpose of generating genetically superior individuals. Therefore, statistical methods are used which allow predict the genetic value of individuals. Recently, a new class of methodologies based on artificial intelligence has been developed and applied to the most diverse areas of Forest Sciences. Artificial Neural Networks (ANNs) have been used in forest modeling in order to estimate the performance of individuals based on traits such as tree diameter, plant height, and volume. Among the ANNs, those called Deep Learning ANN have been used to perform complex functions and solve problems. The use of such technologies has occurred concomitantly with traditional techniques, although some authors have advocated their replacement, given in some cases the best results provided. However, details of their operating mechanisms still remain a "black box". In this sense, this project aims to investigate the use of the ANN technique seeking the practical application of these methods in the prediction of genetic value, when applied in E. grandis and E. urophylla hybrid progeny test. We also propose to perform a comparative evaluation with the REML-BLUP method, widely used in forest genetics.(AU)