1. Root imaging by Minirhizotron

1.1 Introduction to Minirhizotron system

Minirhizotron (MR) system is an non-destructive tool to study roots. MR system is made up of durable, transparent tubes installed in the ground and a light-weight cylindrical imaging device which is lowered into the tubes to collect high-resolution images and track root growth season to season.

1.2 MR image name

Each MR image is given an exclusive name by minirhizotron system (manufactured by Bartz). This name contains important information we need for later analysis.

The dimension of each image is 13.5 mm (width) * 18.0mm (length). Each image is similar to the size of thumb nail. That’s why this system called Minirhizotron.

Location number is the critical information for root study. From this we can calculate the depth roots appear, and root distribution pattern throughout the soil profile.

Note: The soil surface location is not necessary starts from window no.1. Sometime when installing MR observation tubes in the field, a long part of them are exposed above-ground due to the hardpan. Therefore camera sits high above the soil. This makes first several images are still above the soil surface. \[ Rooting\ depth = (Deepest\ root\ window\ no. -\ real\ soil\ surface\ window\ no.) \times 13.5\ mm \]

MR system allows us to observe root dynamics since we can observe the same root repeatedly. This helps us to uncover the life story of each single root.

2. Anotate Minirhizotron images

For root annotation , a open source software called Rootfly is used. This tool is specially designed for MR images. It reads the image name, then arrange images according to the order of tube, window(i.e., location), and session. Values of root length and root diameter are output.

**Rootfly**

Rootfly

**Rootfly workspace**

Rootfly workspace

3. Analyze root data with R

3.1 Workflow

Followed is the workflow diagrams of analyzing Minirhizotron data. From the 4th step onward, can all be processed in R.

3.2 Raw data

The data exported from Rootfly is in .csv file (1_pepper_mr_2021.csv). Keep the raw data raw is a good habit in data analysis. Therefore, we copy this file and save it as a .xlsx file (2_pepper_mr_2021.xlsx) since .csv file does not save figures).

Column names are given by Rootfly software. Length date(1) meaning root length measured at the 1st session. Critical columns include Tube, Window, Length data, Diameter data.

3.3 Process the data

In the Hands-on MR data analysis withR in Github, you will find three R scripts. We will process MR data step by step:

  1. In 1_pepper_mr_2021_prepare.R, raw data is loaded and write in a new excel workbook named 5_pepper_prepared.xlsx, then add new columns of treatment id (3_pepper_treatmentid.xlsx) and soil surface window number (4_pepper_surfaceid.xlsx). The real rooting window number is calculated accordingly then saved together with other selected columns in the second worksheet.

  2. In 2_pepper_mr_2021_pivot.R, data are summerized according to the need (e.g. calculate total root length and average root diameter). This step is similar as the process of generate Pivot table in Excel.

In Excel, the pivot table will look like this:

 

In R, for example, followed chunk of code will give the same result as Pivot,

pepper_mr_pivot_3 <- pepper_mr_sorteddata %>% 
  filter(window_real <= 21 & window_real >10) %>% 
  group_by(variety, plant_no) %>%
  summarize(TRL1 = sum(`Length date(1)`, na.rm = TRUE), 
            TRL2 = sum(`Length date(2)`, na.rm = TRUE),
            RD1 = mean(`Diameter date(1)`, na.rm = TRUE),
            RD2 = mean(`Diameter date(2)`, na.rm = TRUE)
            )
  1. In 3_pepper_mr_2021_ttest.R, data are checked for the assumption of t-test. Then we make graphs with significance and export graph to a PPT. Followed is one of the graphs.