Deep Phenomics: Advanced image-based trait discovery in Megathyrsus maximus for enhanced forage breeding

• Grant number: 25/20969-3
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate
• Start date: March 01, 2026
• End date: August 31, 2026
• Field of knowledge: Biological Sciences - Genetics - Quantitative Genetics
• Principal Investigator: Anete Pereira de Souza
• Grantee: Guilherme Francio Niederauer
• Supervisor: Ricardo da Silva Torres
• Host Institution: Centro de Biologia Molecular e Engenharia Genética (CBMEG). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Institution abroad: Wageningen University & Research,
• Associated to the scholarship: 23/06910-0 - Unraveling molecular mechanisms of quantitative traits in Megathyrsus maximus using digital phenotypes and computer vision, BP.DR

Abstract

Tropical forages are essential for Brazilian livestock farming and, consequently, for the national economy. However, their genetic improvement programs face considerable challenges. Progress is hindered by the limited availability of genomic resources, the complexity of polyploidy, and the challenges of phenotyping large Megathyrsus maximus(guinea grass) populations. Although traditional strategies for evaluating traits of interest, such as leaf dry matter production, regrowth, and pathogen resistance, are well-established, they are inefficient, requiring substantial time, high costs, and significant manual effort. To overcome these limitations, high-throughput phenotyping techniques, including those based on imaging and remote sensing, are emerging as modern solutions. However, the substantial volume of data generated by these technologies necessitates the development of novel analytical approaches. In this context, this project aims to develop an innovative methodology, grounded in deep learning, to extract complex phenotypes from remote sensing images (RGB and multispectral) of a M. maximus germplasm population. The internship at Wageningen University \& Research (WUR) will focus on image pre-processing and the development of convolutional neural networks (CNNs) for automated segmentation of regions of interest and acquiring novel digital phenotypes. These phenotypes will be integrated into modern molecular breeding approaches, such as genome-wide association studies (GWAS), later in the Ph.D. project. This work offers new opportunities to improve phenotyping efficiency and, consequently, accelerate the development of improved forage cultivars. (AU)