MyROOT: A novel method and software for the semi-automatic measurement of plant root length

Abstract

28Root analysis is essential for both academic and agricultural research. Despite the great 29 advances in root phenotyping and imaging however, calculating root length is still 30 performed manually and involves considerable amounts of labor and time. To overcome 31 these limitations, we have developed MyROOT, a novel software for the semi-automatic 32 quantification of root growth of seedlings growing directly in agar plates. Our method 33 automatically determines the scale from the image of the plate, and subsequently 34 measures the root length of the individual plants. To this aim, MyROOT combines a 35 bottom-up root tracking approach with a hypocotyl detection algorithm. At the same time 36as providing accurate root measurements, MyROOT also significantly minimizes the user 37 intervention required during the process. Using Arabidopsis, we tested MyROOT with 38 seedlings from different growth stages. Upon comparing the data obtained using this 39 software with that of manual root measurements, we found that there are no significant 40 differences (t-test, p-value < 0.05). Thus, MyROOT will be of great aid to the plant 41 science community by permitting high-throughput root length measurements while 42 saving on both labor and time. 43 44 48 impact of roots in agriculture 1 . As such, generating tools for precise, high-throughput 49 phenotyping and imaging of the root is essential for plant research and agriculture. Even 50 phenotyping facilities such as the ones available in the European Plant Phenotypic 51 Network (http://www.plant-phenotyping-network.eu/) have started to implement tools for 52 the massive screening of roots.53Roots provide the necessary structural and functional support for the incorporation of 54 nutrients and water from the soil. In Arabidopsis thaliana (Arabidopsis), the primary root 55 has a very simplified anatomy that makes it very amenable for genetic and microscopic 56 analyses 2-4 . Different root cell lineages are derived from the activity of a group of stem 57 cells located at the root apex. Here, the stem cell niche is formed by a few (3-7) quiescent 58 center (QC) cells that occasionally divide asymmetrically to renew themselves and to 59 3 form daughter stem cells. From the root apex, these cells actively divide in the 60 meristematic zone, and before exiting the cell cycle in the transition zone, continue to 61 elongate and differentiate in spatially separated regions of the root. In this way, primary 62 root growth is determined by the balance between cell division and cell elongation within 63 the different zones of the root 5-8 . 64The most straightforward symptom of abnormal root growth or development can be 65 identified by examining the length of the primary root in seedlings. Abnormalities in 66 length can usually be observed and measured just five to six days after germination 67 (DAG), where still reflect their embryonic origin 9 . Growth defects in the primary root of 68 seedlings are not only consistent with overall growth defects, but also persistent along the 6...

Recent advances in methods for in situ root phenotyping

“…Currently, there are many semi-automated root analysis software, including EZ-Rhizo ( Armengaud et al, 2009 ), GrowScreen-Root ( Nagel et al, 2009 ), GiA Roots ( Galkovskyi et al, 2012 ), GLO-RIA ( Rubén et al, 2015 ), KineRoot ( Basu et al, 2007 ), MyROOT ( Betegón-Putze et al, 2018 ), Multi-ADAPT ( Ishikawa & Evans, 2010 ), RootNav ( Pound et al, 2013 ), RootReader2D ( Clark et al, 2013 ), RootScape ( Ristova et al, 2013 ), RootTipTrace ( Geng et al, 2013 ), and SmartRoot ( Lobet, Pagès & Draye, 2011 ). EZ-Rhizo is a Windows-integrated and semi-automated computer program that can be employed to quantify multiple root parameters of plants growing on agar medium.…”

Root anatomy based on root cross-section image analysis with deep learning

“…Other investigations have focused on identifying plant diseases [34] or on stress phenotyping [26]. Furthermore, several prior studies have focused on data-driven approaches and tools for belowground plant phenotyping, including identifying and quantifying root morphological parameters, such as changes in root architecture, or branching and growth [35,36,37,38]. Such approaches rely on standard image analysis techniques as opposed to state-of-the-art deep learning.…”