Root interact with the soil environment in complex ways. We propose a series of tools to mine for patterns in the 3D trajectories of roots.



We propose an ImageJ plugin for direct fitting of root trajectory on a series of tomographic projection data termed MultiplaneTracking. The algorithm is first initiated with a corse tracing that has either been done manually or with existing tools. We used the the Vessel tracking from MeVisal



The root centrelines were analysed as analogue 3 dimensional signals to extract both the frequency and radius of the Helical patterns present in the root trajectory. First, the background noise of the signal was obtained using a spline regression with 3 anchor points. The background noise was then removed from the raw signal, which resulted in root tracings being centred along the z-axis (Supplementary Information). Centred signals were then analysed for periodic features, and in particular for helical periodicity. The analysis was based on a modification of the Fourier transform that use a set of orthonormal helix forming basis functions. A custom software tool termed RootHix was developed to perform such tasks. Software tools for analysis of root centreline are freely available on GitHub The repository contains also sample files of lentils roots growing in transparent soil under vairous levels of compression

Detailed information on the mathematical transform used to extract the coefficients of the helix can be found here.



Phosphatase activity tracer

Zymography is a powerful techniques to quantify phosphatase activity along the root using quantitative imaging. This page provides an imageJ plugin that extract the profiles of phosphatase activity along the root. The plugin is initiated with a coarse outline of the root centerline. An algorithm is then used to refine the position of the centerline along the root. The profile of activity is then extracted perpendicular to the centerline and exported as a CSV file. Python scripts are also provided to post-proces the data. scripts consists of an optimisation algorithm to find the enter of the root more accurately but also the fitting of models to extract key features of the profile.

This file is the ImageJ plugin (.java)

This file is the Python script to proces data extracted from images (.py)


This page provides softwarefor model based extraction of phenotyping data. The files attach are python codes for:

1. Construction of kernel based root length density functions

2. Determination of the optimal smoothing parameter for the kernel estimator

3. Simulation of Meristematic wave models with time detay and root types

3. Optimisation algorithm for determination of growth prameters from data

The code and software is made available on GitHub

ImageJ plugin for identification of QR code can be downloaded here

ImageJ plugin for identification of growth zone can be downloaded here

ImageJ plugin for root tracing can be downloaded here

Please contact me for instruction on installation and usage of the tools provided here. 

THese tools are provided as is, without warranty etc...


Video S1: Simulation of a root tip entering a patch of bacteria using Smooth Particle Hydrodynamic (SPH) approach. Attached bacteria on the root surface (red circle) move slowly away from the tip, and this results in an accumulation of bacteria at the root tip observed experimentally. Free bacteria (blue circle) in soil however move away from the tip at constant velocity. Free bacteria grow less because they can’t remain at the root tip. Graphs below the rhizosphere model indicate bacterial density for attached bacteria (red) free moving bacteria (blue) and carbon concentration (green).


EffectOfAttachment short

Video S2: Effect of microbial attachment bacteria density. The effect of attachment can be visualised by activating attachment once the steady state without attachment is observed. This produce a drastic increase in the density of attached bacteria with a very small decrease in the density of free moving bacteria. Overall, the total density of bacteria in the root tip is increased 

 without cap        with cap

Video S3: Bacterial density in response to exposure to bacteria during a short period of time. In the absence of root cap (left) all bacteria disappear. When a root cap is acting as a reservoir (right) bacterial density maintains in low quantity at the root tip



Video S4: Simulation of a root tip entering a patch of bacteria with high chemotactic coefficient. At first contact with the root, bacterial velocity is lower than root tip velocity and bacteria are displaced away from the tip. However, after about 60 to 80 h spent in contact with the root tip, the bacterial population attain a critical size and progress towards the root tip. A bacterial front is formed and its profile is similar to a supersonic booms. At steady state, the front is placed in the elongation zone and corresponds to the peak of carbon density. Graphs below the rhizosphere model indicate the equivalent bacterial density for free moving bacteria (blue) and carbon concentration (green).