DIRT-Pop is a robust pipeline tailored for clustering root architecture types. It's specially designed to work with outputs produced by the Digital Image of Root Traits (DIRT) software.

Before running DIRT-Pop, ensure the following software are installed:

• Docker
• Git

Clone the DIRT-Pop repository to your local machine:

git clone https://github.com/tomatopepper/DIRT-Pop.

Once cloned, navigate into the repository's directory:

cd DIRT-Pop

Initiate the DIRT-Pop pipeline using Docker:

docker run -it -v $(pwd):/opt/DIRT-Pop computationalplantscience/dirtclust bash

Finally, execute the DIRT-Pop clustering script:

N_TRIAL_KMEANS=<number_of_clustering_trials> \
N_TRIAL_KNEEDLE=<number_of_kneedle_trials> \
N_CLUSTER=<number_of_clusters> \
INPUT=/opt/DIRT-Pop/DIRTtest.csv \ # your DIRT file 
WORKDIR=/opt/DIRT-Pop ./wrapper.sh

Replace the following placeholders with appropriate values:

• <number_of_clustering_trials>: Specify the number of clustering iterations (e.g., 25).
• <number_of_kneedle_trials>: Specify the number of kneedle algorithm tests (e.g., 25).
• <number_of_clusters>: Specify the max number of clusters you want to look for (e.g., 25 means searching for 1-25 clusters)
• Note: Modify the INPUT path if your DIRT output file is located elsewhere or named differently.

For any inquiries or issues, please contact [email protected].

We thank Wes Bonelli for helping create docker environment for DIRT-Pop pipeline.