Skip to main content

Soybean Root Studies

Root phenotyping pipeline
Root phenotyping pipeline (Falk et al. 2019)

We leverage computer vision and machine learning methods to develop root trait phenotyping pipelines. Recently, we used this pipeline to conduct a large-scale root system architecture (RSA) traits examination with time series data on 292 diverse soybean accessions. RSA traits are tedious to phenotype, which suffer from the limitation of scale and scope, and are susceptible to measurement variation. We combined 50K single nucleotide polymorphism (SNP) markers with our phenomics platform to decipher the genetic diversity and explore informative root categories based on current literature for root shape categories, for example drought tolerance, nutrient acquisition. We used marker-based and phenotype-based clusters to examine trait diversity and correlate the diversity with descriptors (metadata). This visualization connects genotypic and phenotypic information to identify useful and unique accessions for breeding. Through the integration of convolution neural network(s) and Fourier transformation methods, we developed methods to capture shape-based clusters that are a novel way for trait cataloging for breeding and research applications. We used the phenotyping platform and new computer analytic capability to identify genes controlling important root-related traits. We are now pursuing further research to validate these genes, and conduct large scale genomic studies. We have also developed software program for automated root nodule phenotyping. New breeding initiatives include the development of soybean varieties selected for root traits in addition to conventional above ground traits.
Outcomes: (1) developed a low-cost higher-throughput, non-destructive HTP platform enabling temporal assessments of root samples, (2) created an ML enabled semi-automated computer-vision program for image capture and curation, (3) published advanced open-source software program that allows the extraction of a multitude of root system architecture traits. We have developed an end-to-end pipeline with a fully automated software package including tunable image thresholding and image-based trait extraction.

Primary collaborators: Dr. Baskar Ganapathysubramanian, Dr. Soumik Sarkar, Dr. Arti Singh.
Funding: Iowa Soybean Research Center, Plant Sciences Institute, R F Baker Center for Plant Breeding, Iowa Soybean Association, Monsanto Chair in Soybean Breeding.

Related Publications:

  • Falk KG, TZ Jubery, JA O’Rourke, A Singh, S Sarkar, B Ganapathysubramanian, AK Singh. 2020. Soybean root system architecture traits study through genotypic, phenotypic and shape based clusters. Plant Phenomics. Accepted (in press).
  • Falk KG, T Jubery, SV Mirnezami, KA Parmley, S Sarkar, A Singh, B Ganapathysubramanian, AK Singh. 2019. Computer Vision and Machine Learning Enabled Soybean Root Phenotyping Pipeline. BMC Plant Methods volume 16, Article number: 5.