Location: Plant Science Research
Title: RootScan: Software for high-throughput analysis of root anatomical traits Authors
|Williams, Michael -|
|Lynch, Jonathan -|
|Brown, Kathleen -|
Submitted to: Plant and Soil
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: January 12, 2012
Publication Date: February 23, 2012
Citation: Burton, A.L., Williams, M.S., Lynch, J.P., Brown, K.M. RootScan: Software for high-throughput analysis of root anatomical traits. Plant and Soil. 357:189-203. 2012. Interpretive Summary: An organism’s phenotype is composed of its observable characteristics, including morphological, physiological, and developmental traits. Phenotype is determined by an organism’s genetic sequence, as well as its environment. Anatomical root traits (or ‘phenes’) influence physiological processes in plants, including growth and resource acquisition. Traditionally, root traits have represented an underutilized area in plant breeding, due to the complex effects of the environment on their growth, and the fact that in situ observation of root systems is challenging. Phenotypic screening is used to identify complex combinations of traits that collectively hold agronomic value. Such data can be used to screen large numbers of genotypes for traits of interest, to characterize genetic variability within a given population, and to identify genetic regions that control a particular trait. The ability to make accurate and rapid measurements of root traits could lead to their use as a focus in plant breeding. However, the development of high throughput phenotyping techniques represents a general limitation to trait-based breeding. Given the inadequate attention given to root traits, this is particularly true for root phenotyping. This manuscript describes new computer software developed for quantitative evaluation of root phenotypic traits.
Technical Abstract: RootScan is a program for semi-automated image analysis of anatomical phenes in root cross-sections. RootScan uses pixel value thresholds to separate the cross-section from its background and to visually dissect it into tissue regions. Area measurements and object counts are performed within various regions of interest, including the stele and cortex. Interim steps in the software remove background debris, and account for variability in light intensity, contrast, edge shadows and image quality. A graphical user interface (GUI) permits the user to see which regions are selected, to edit those selections, and to rate and comment on the data output. The structure of the program allows for organized workflow and increased data collection efficiency. The program collects data on more than 20 variables per image including areas of the cross-section, stele, cortex, aerenchyma lacunae, xylem vessels, and counts of cortical cells and cell files. An increased rate of data collection more than doubles the number of images that can be analyzed per hour over comparable methods, and allows collection of four times the number of variables. Correlation analysis shows that data from RootScan is equal or greater in accuracy than data collected with Photoshop. Compared with currently available tools, this software offers considerable improvements in the amount and quality of data generated, and the time needed for data collection. RootScan permits phenotypic scoring of physiologically and agronomically important traits on a large number of genotypes.