Unraveling molecular mechanisms of quantitative traits in Megathyrsus maximus using digital phenotypes and computer vision

Abstract

Tropical forage plants are cultivated crops that play a significant role in the feed supply for Brazilian livestock and consequently contribute to the country's economy. Despite the importance of this crop, the efficiency of its breeding programs, which are relatively recent, requires further enhancements. The limited genomic resources available and the absence of advanced methodologies for species such as Megathyrsus maximus hinder the overcoming of existing obstacles, which involve aspects such as chromosomal polyploidy and reproduction through apomixis. By applying high-throughput phenotyping techniques, the feasibility of acquiring different phenotypic measurements using remotely sensed images will be determined. Computer vision approaches will be employed to extract morpho-physiological characteristics from the images, enabling the digitized definition of various phenotypic traits. These digital phenotypes will be associated with genotyping data through genomic association models, allowing for the identification of correspondences between the defined phenotypes and observed genotypic differences using genetic markers (SNPs). In this context, the genetic potential of individuals within the target population will be estimated, and the results will be used for the selection of superior genotypes and comparison with traditional phenotyping and associative modeling techniques. Finally, RNA-Seq techniques will be employed to evaluate differential gene expression among contrasting genotypes, thus elucidating the molecular mechanisms involved in the control of the regrowth process, which is of great economic interest in this species.It is expected that the integration of the proposed approaches will lead to the development of a methodological framework for genomic prediction, incorporating more robust and sophisticated phenotyping and genotyping methodologies. In addition to enabling a more efficient ranking of genotypes for genomic selection, this framework is expected to provide a methodological tool for inferring new genomic associations, serving as a basis for advanced biotechnology strategies. The project has the potential to introduce new practices to the Brazilian agricultural sector, thereby improving the productivity, quality, and sustainability of beef cattle farming. (AU)