Uncovering the genetic architecture of stink bug resistance and seed potential in soybean through image-based phenotyping and QTL mapping

• Ano de defesa: 2025
• Autor(a) principal: Braga, Patricia
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Fenotipagem de alto rendimento; High-throughput phenotyping; Insect resistance; Mapeamento QTL; Melhoramento de precisão; Potencial fisiológico de sementes; Precision breeding; QTL mapping; Resistência a insetos; Seeds physiological potential
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/11/11137/tde-05022026-155222/

Stink bugs are among the most critical factors compromising soybean productivity, causing substantial reductions in yield and seed quality. This study combined image-based phenotyping, conventional field evaluations, genotyping-by-sequencing (GBS), and QTL mapping to investigate the genetic architecture related to stink bug resistance and seed physiological potential in soybean. A population of recombinant inbred lines (RILs), derived from the cross between the resistant cultivar IAC-100 and the susceptible cultivar CD-215, was analyzed using classical manual phenotyping measures and RGB, hyperspectral, and X-ray imaging across two growing seasons (2018/2019 and 2019/2020). Advanced imaging techniques provided detailed insights into the physical, chemical, and morphological attributes of seeds, detecting subtle differences and highlighting stink bug-induced damage. The integration of image-derived data with traditional field evaluations enabled high-accuracy predictions of traits such as healthy seed weight (>90%) and stink bug tolerance (60%). Multivariate analyses revealed three main trait clusters: stink bug resistance, seed shape, and seed physiological potential. Quantitative trait loci (QTL) mapping and machine learning approaches identified significant loci associated with resistance and seed quality, distributed across 10 chromosomes, particularly on chromosomes 1, 12, and 17. Functional annotations revealed enriched pathways related to oxidative stress response and secondary metabolism, reinforcing the biological relevance of the findings. High heritability values for traits such as days to maturity (87%) and tolerance (80%) underscore the robustness of the phenotyping approach. These results demonstrate the potential of integrating high-throughput phenotyping and genomic tools to accelerate the development of elite soybean cultivars with enhanced biotic stress resistance and superior agronomic performance. Future advancements in image acquisition automation will further expand the applicability of these methods to breeding program routines.