Unlocking the power of AI for phenotyping fruit morphology in Arabidopsis
Deep learning can revolutionise high-throughput image-based phenotyping by automating the measurement of complex traits, a task that is often la bour-intensi v e, time-consuming, and prone to human err or. Howev er, its pr ecision and adapta bility in accuratel y phenotyping organ-level tr aits, suc...
| Autores principales: | , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | Inglés |
| Publicado: |
Oxford University Press
2025
|
| Materias: | |
| Acceso en línea: | https://academic.oup.com/gigascience/article/doi/10.1093/gigascience/giae123/8010444 http://hdl.handle.net/20.500.12324/41152 https://doi.org/10.1093/gigascience/giae123 |
| Sumario: | Deep learning can revolutionise high-throughput image-based phenotyping by automating the measurement of complex traits, a task that is often la bour-intensi v e, time-consuming, and prone to human err or. Howev er, its pr ecision and adapta bility in accuratel y phenotyping organ-level tr aits, suc h as fruit morphology, remain to be fully evaluated. Establishing the links between phenotypic and genotypic variation is essential for uncovering the genetic basis of traits and can also provide an orthologous test of pipeline effecti v eness. In this study, we assess the efficacy of deep learning for measuring variation in fruit morphology in Arabidopsis using ima ges fr om a m ultipar ent adv anced gener ation inter cr oss (MAGIC) mapping famil y. We tr ained an instance se gmentation model and developed a pipeline to phenotype Arabidopsis fruit morphology, based on the model outputs. Our model achieved strong perfor- mance with an av era ge pr ecision of 88.0% for detection and 55.9% for segmentation. Quantitati v e trait locus analysis of the derived phenotypic metrics of the MAGIC population identified significant loci associated with fruit morphology. This analysis, based on au- tomated phenotyping of 332,194 individual fruits, underscores the capability of deep learning as a robust tool for phenotyping large populations. Our pipeline for quantifying pod morphological traits is scala b le and provides high-quality phenotype data, facilitating genetic analysis and gene discov er y, as well as advancing crop breeding resear c h. |
|---|