Update:
Project Summary:
Research Goals: To develop superior soybean cultivars for Iowa farmers. These cultivars will provide high yield and protection against biotic and abiotic factors to keep Iowa farmers competitive and profitable.
Our long term team goal is to build highly desirable trait packages for our Iowa farmers. These will include an arsenal of performance genes and resistance against insects and diseases in soybean. Our efforts to integrate performance using high throughput phenotyping (ground and aerial) will help to continue national and international prominence of our breeding and research activities and lead to benefits to Iowa farmers, farm economy and industries.

Project Objectives:
The objectives of this project are:
1. Increase soybean seed yield using genomic and phenomic tools.
2. Incorporate high yield and multiple yield protection traits using improved germplasm material.
3. Establish and operate a marker-assisted selection breeding system and repository.
4. Publish results from our current year funding on (a) determination of plant traits in soybean for maximizing profitability in corn-soybean rotation, and (b) seed density and row width studies.

Project Deliverable:
Short term: development of breeding populations for objectives listed in this project.
Long term: cultivar and germplasm lines that meet the requirements of Iowa farmers and industry of high yield trait with yield preservation.

Proposed specific deliverable (long term; 2020): IA1022 (SCN), IA2102 and IA3048 (SCN) are used as commercial checks in the cooperative uniform testing for MG I, II and III, respectively. We will take the mean of top five yielding experimental genotypes in each maturity group (I, II and III) and propose a 1% yield gain per year over the mean of those top 5 ranking genotypes compared to these checks (5 year goal). Additionally, starting in 2019, we will include commercial checks in our state wide testing enabling us to demonstrate performance of our lines relative to current commercial checks.

Benefit to Soybean Farmers:
Our team is working to provide genetic solutions to Iowa farmers that meet their on-farm requirements and help improve profitability. This project will lead to development of cultivars that addresses their on-farm production issues and requirements; and the development of unique germplasm that will be desirable to private industry and indirectly benefit Iowa farmers. Market ready cultivars will also be another opportunity for IA farmers, who will have a choice to grow ISU developed cultivars. This project will build on existing breeding network at ISU. This partnership will also be beneficial for establishing a strong foundation to the Iowa State University’s Soybean Research Center, United Soybean Board, Seed Companies, USDA-NIFA and other agencies that support specific research projects.
The ISA funding is enabling graduate students (Kyle Parmley, Kevin Falk, John Shook, Matt Carroll, Clayton Carley) to work on cutting edge technology driven breeding projects integrating phenomics and genomics in breeding. Our students are being actively recruited by private seed company for breeder positions. Graduate students (4) are employed as plant breeders. This is an excellent return on investment as the next generation breeders in leading companies are coming from ISA funded breeding program. These students are passionate and committed to IA farmers and their on-farm issues that can be solved through science and technology.


Progress Report:

Highlights of research project outcomes:
* Root related research: Root system architecture (RSA) traits are of interest for breeding selection; however, measurement of these traits is difficult, resource intensive, and results in large variability. The advent of computer vision and machine learning (ML) enabled trait extraction and measurement has renewed interest in utilizing RSA traits for genetic enhancement to develop more robust and resilient crop cultivars. In a research articles (published in Plant Methods), we report a mobile, low-cost, and high-resolution root phenotyping system developed in-house, and composed of an imaging platform with computer vision and ML based segmentation approach to establish a seamless end-to-end pipeline - from obtaining large quantities of root samples through image based trait processing and analysis. This high throughput phenotyping system, can handle hundreds to thousands of plant root phenotyping samples, integrates time series image capture coupled with automated image processing that uses optical character recognition (OCR) to identify seedlings via barcode, followed by robust segmentation integrating convolutional auto-encoder (CAE) method prior to feature extraction. The pipeline includes an updated and customized version of the Automatic Root Imaging Analysis (ARIA) root phenotyping software. Using this system, we studied diverse soybean accessions from a wide geographical distribution and genetic variability for RSA traits, including root shape, length, number, mass, and angle. This system provides a high-throughput, cost effective, non-destructive methodology that delivers biologically relevant time-series data on root growth and development for phenomics, genomics, and plant breeding applications. This phenotyping platform is designed to quantify root traits and rank genotypes in a common environment thereby serving as a selection tool for use in plant breeding.
Ooutcome for the program: We initiated breeding crossed to include root traits in our soybena breeding efforts. This is an on-going work.
Relevant paper: 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, 16 (1), 5

* Yield related research: We proposed the concept of prescriptive plant breeding driven by machine learning. This work was published in Scientific Reports) where we explored the capability of fusing high dimensional phenotypic trait (phenomic) data with a machine learning (ML) approach to provide plant breeders the tools to do both in-season seed yield (SY) prediction and prescriptive cultivar development for targeted agro-management practices (e.g., row spacing and seeding density). Phenotypic trait data (canopy temperature, chlorophyll content, hyperspectral reflectance, leaf area index, and light interception) were measured using an array of sensors at growth stages during the growing season and SY determined by machine harvest. Random forest (RF) was used to train models for SY prediction using phenotypic traits (predictor variables) to identify the optimal temporal combination of variables to maximize accuracy and resource allocation. RF models were trained using data from both experiments and individually for each agro-management treatment. We report the most important traits agnostic of agro-management practices. Several predictor variables showed conditional importance dependent on the agro-management system. We assembled predictive models to enable in-season SY prediction, enabling the development of a framework to integrate phenomics information with powerful ML for prediction enabled prescriptive plant breeding.
Relevant papers: Parmley KA, RH Higgins, B Ganapathysubramanian, S Sarkar, AK Singh. 2019. Machine Learning Approaches for Prescriptive Plant Breeding. Scientific Reports. v9, Article number: 17132
* Yield related research: Phenomic assisted breeding (PAB) methodologies have lagged those of genomic-assisted techniques with a slower rate of advancement; although PAB is now a critical component of mainstream cultivar development pipelines. Advancements made in phenotyping technologies have empowered plant scientists with affordable high-dimensional datasets to optimize the operational efficiencies of breeding programs. Machine learning method was used to map complex relationships between phenomic traits and seed yield and prediction performance assessed using two cross-validation (CV) scenarios consistent with breeding challenges. To develop a prescriptive sensor package for future high-throughput phenotyping deployment to meet breeding objectives, feature importance in tandem with a genetic algorithm (GA) technique allowed selection of a subset of phenotypic traits, specifically optimal wavebands. The results illuminated the capability of fusing ML and optimization techniques to identify a suite of in-season phenomic traits that will allow breeding programs to decrease the dependence on resource-intensive end-season phenotyping (e.g., seed yield harvest). While we illustrate with soybean, this study establishes a template for deploying multi-trait phenomic prediction that is easily amendable to any crop species and any breeding objective.
Parmley K, K Nagasubramanian, S Sarkar, B Ganapathysubramanian, AK Singh. 2019. Development of Optimized Phenomic Predictors for Efficient Plant Breeding Decisions using Phenomics Assisted Selection in Soybean Plant Phenomics. Science Plant Phenomics. vol. 2019, Article ID 5809404.
Outcome for the program: We are building on this successful project in our breeding program to develop cultivars. This is an on-going work.

2018-19 Progress report (ISU soybean breeding and research program)
Breeding Success:
• Eighteen new soybean varieties were developed in 2018 (current funding cycle) and invention disclosures were submitted to ISU research foundation. These 18 varieties are undergoing final year of statewide performance testing in 2019, prior to commercial seed sales.
• 11 varieties are food-grade soybean providing estimated $50-100 of net return per acre (Data source: ISURF); 7 varieties are suitable for traditional (non-GM) markets providing opportunities for IA farmers.
• Six varieties have >50% oleic acid and one has over 75% oleic acid for specialty oil markets.
• In late MG I, varieties with seed yield 34-45% higher than check IA1022 were developed.
• In mid MG II, varieties with seed yield up to 6% higher than check IA2102 were developed.
• In early MG III, varieties with seed yield up to 13% higher than IA2102 and 5% higher than Uniform test check LD11-2170 check were developed.
• In 2018, IA20023 (non-GM, food grade variety) out yield several herbicide trait commercial seed company varieties showing that ISA funded project varieties from ISU are yield competitive with larger companies [Data source: Iowa Crop Improvement Association variety performance test results].

Student success:
• Four PhD student have graduated from the program and three are working as breeders: Race Higgins (PanAm seeds) Kyle Parmley (Bayer), Kevin Falk (Corteva).
• Kevin Falk won several national and institutional recognition: department of Agronomy Outstanding Graduate Student Award, ISU’s graduate and professional student senate (GPSS) research award, ISU’s GPSS leadership award, National Association of Plant Breeders Borlaug Scholar, and American Seed Traders Association Student Video – Grand Prize (which aimed to debunk myths around seed industry).
• Kyle Parmley was selected for the ISU graduate college emerging leader academy, and won several travel awards to make scientific presentations. Kyle was the only invited graduate student speaker at the 2018 R F Baker Plant Breeding symposium at ISU.
• Three of our current PhD students (John Shook, Matt Carroll, Clayton Carley) were selected for National Research Traineeship in Predictive Plant Phenomics at ISU. Student have made presentation nationally and traveled internationally

View uploaded report

The Iowa Soybean Association funding enabled us to develop new soybean varieties with a combination of high yield and favorable production and quality traits, of which seven will undergo seed increase and certification for commercial launch in 2021 and onwards. Innovative plant breeding related concepts were proposed and validated including phenomic prediction of yield and other traits, prescriptive plant breeding catered to develop varieties and their placement to meet the need of farmers, and building cultivars with better root traits diversity to increase yield production. Our team developed and used high throughput phenotyping for breeding and research (aerial and ground), with operationalization of UAVs and camera systems. Thirteen peer reviewed papers were published in 2018-2019. Two graduate students completed their PhD and started working in the industry to continue assisting farmers through innovative and cost effective solutions. Our team organized the 2nd International Workshop on Machine Learning for Cyber-Agricultural Systems in Ames, IA (2019). Several tours were hosted including for farmers, industry and students to show plant breeding and research pipeline.