A growth phenotyping pipeline for Arabidopsis thaliana integrating image analysis and rosette area modeling for robust quantification of genotype effects
- To gain a deeper understanding of the mechanisms behind biomass accumulation, it is important to study plant growth behavior. Manually phenotyping large sets of plants requires important human resources and expertise and is typically not feasible for detection of weak growth phenotypes. Here, we established an automated growth phenotyping pipeline for Arabidopsis thaliana to aid researchers in comparing growth behaviors of different genotypes. The analysis pipeline includes automated image analysis of two-dimensional digital plant images and evaluation of manually annotated information of growth stages. It employs linear mixed-effects models to quantify genotype effects on total rosette area and relative leaf growth rate (RLGR) and ANOVAs to quantify effects on developmental times. Using the system, a single researcher can phenotype up to 7000 plants d(-1). Technical variance is very low (typically < 2%). We show quantitative results for the growth-impaired starch-excessmutant sex4-3 and the growth-enhancedmutant grf9. We showTo gain a deeper understanding of the mechanisms behind biomass accumulation, it is important to study plant growth behavior. Manually phenotyping large sets of plants requires important human resources and expertise and is typically not feasible for detection of weak growth phenotypes. Here, we established an automated growth phenotyping pipeline for Arabidopsis thaliana to aid researchers in comparing growth behaviors of different genotypes. The analysis pipeline includes automated image analysis of two-dimensional digital plant images and evaluation of manually annotated information of growth stages. It employs linear mixed-effects models to quantify genotype effects on total rosette area and relative leaf growth rate (RLGR) and ANOVAs to quantify effects on developmental times. Using the system, a single researcher can phenotype up to 7000 plants d(-1). Technical variance is very low (typically < 2%). We show quantitative results for the growth-impaired starch-excessmutant sex4-3 and the growth-enhancedmutant grf9. We show that recordings of environmental and developmental variables reduce noise levels in the phenotyping datasets significantly and that careful examination of predictor variables (such as d after sowing or germination) is crucial to avoid exaggerations of recorded phenotypes and thus biased conclusions.…
Verfasserangaben: | Samuel Janne ArvidssonGND, Paulino Perez-Rodriguez, Bernd Müller-RöberORCiDGND |
---|---|
DOI: | https://doi.org/10.1111/j.1469-8137.2011.03756.x |
ISSN: | 0028-646X |
Titel des übergeordneten Werks (Englisch): | New phytologist : international journal of plant science |
Verlag: | Wiley-Blackwell |
Verlagsort: | Malden |
Publikationstyp: | Wissenschaftlicher Artikel |
Sprache: | Englisch |
Jahr der Erstveröffentlichung: | 2011 |
Erscheinungsjahr: | 2011 |
Datum der Freischaltung: | 26.03.2017 |
Freies Schlagwort / Tag: | development; growth; leaf area; modeling; phenotyping |
Band: | 191 |
Ausgabe: | 3 |
Seitenanzahl: | 13 |
Erste Seite: | 895 |
Letzte Seite: | 907 |
Fördernde Institution: | German Ministry of Research and Education ('GoFORSYS - Potsdam-Golm BMBF Forschungseinrichtung zur Systembiologie, Photosynthesis and Growth) [FKZ 0313924]; EU (Marie Curie Research Training Network 'VaTEP - Vacuolar Transport Equipment for Growth Regulation of Plants' [MRTN-CT-2006-035833]; Federal State of Brandenburg; Max-Planck Institute of Molecular Plant Physiology; Subdireccion de Investigacion: Linea 15, Colegio de Postgraduados, Mexico |
Organisationseinheiten: | Mathematisch-Naturwissenschaftliche Fakultät / Institut für Biochemie und Biologie |
Peer Review: | Referiert |