Harvest for Health Breakthrough Crop Challenge

Challenge – Predictive Model Development and Validation: Applicants will develop and submit a validated predictive model outlined in the Submission Materials section. All applications will be reviewed by an expert review panel and $1 million will be awarded to the selected predictive model that meets all outlined criteria. Applications for the Breakthrough Crop Challenge will be due by August 7, 2024. FFAR and GAIN reserve the right not to issue an award if no predictive model meets all criteria outlined.

Any individual(s) with the skills, knowledge and resources necessary to carry out the Breakthrough Crop Challenge as Program Director(s)/Principal Investigator(s) may apply through their home institution or organization. The Foundation for Food & Agriculture Research welcomes applications from all domestic and international higher education institutions, non-profit and for-profit organizations and U.S. government-affiliated research agencies. Applicants will be required to agree to the Breakthrough Crop Challenge Terms and Conditions and to develop and submit a validated predictive model as outlined in the submission criteria section below. All applications will be reviewed by an expert review panel for eligibility to receive the U.S. $1.0 million award.

No, applicants do not need to have participated in Seed Funding to participate in the Predictive Model Development and Validation or potentially receive $1 million.

One applicant/research team will receive the $1 million award

Yes. All applications must be submitted in English and all award payments will be disbursed in U.S. dollars, and any institution around the globe may apply.

Yes. For-profit institutions may apply to the opportunity.

Underutilized crops are mostly wild or semi-domesticated species adapted to local environments. These crops were used as traditional foods for centuries but became increasingly neglected in farming systems. Agricultural modernization, widespread monoculture and the promotion of high-yielding varieties have marginalized these crops (Li et al., 2020).

Underutilized crops can also be defined by their limited geographic distribution or use, and the absence of their broader adoption or support from policymakers, technology providers, donors, breeders and extension services (Khan et al., 2022).

No, you can select any category of your choice. If high nutrient density is selected as the property of interest, another functional property must also be selected to make the application more competitive. A predictive model that selects more than one non-nutrient functional property will be considered more competitive.

Yes.

There are no limitations on location or type of underutilized crop.

Yes.

Yes.

Input data should be limited to at least one or a combination of the following: genomic sequencing, nutrition composition, and biochemical properties/composition of crops.

FFAR will employ a jury of panels to review the submitted model. This will be a 2-step review process:

Review Step 1
All applicants will submit background information and information on the submission criteria section I-XII. This will undergo peer review and up to 10 applications will be selected for the second review step.

Review Step 2.1
Up to 10 applicants from Review Step 1 will provide the predictive model and detailed instructions on running the predictive model as a one-line executable algorithm, with input and output data for assessment and validation, trained model, and the data set used to train the model with a description of how the training was conducted. An external computational contractor will run the model against an evaluation dataset and will synthesize a performance report. Panel reviewers will review the synthesized report and recommend up to five applications for further wet lab testing of predicted functional property(ies). Selection criteria will be based on performance metrics, such as accuracy, precision, recall and F1-measure.

Review Step 2.2
Up to 5 applications will undergo another predictive model evaluation using a dataset compiled by FFAR and wet lab testing to test the functionality predicted by the model in a food science/food processing lab.

The generated report will be reviewed by the panel to determine a winner for the $1M award.

Overall, an independent contractor/ entity will:

• Test the model’s ability to predict functional properties in selected crops.
• Test the ability of the model to predict functional properties in crops selected by FFAR.
• Test predicted functionality in different solutions/ models/assays or products for potential applicability and prediction accuracy.

The Predictive Model Development and Validation Challenge winner will be involved with FFAR in the utilization of the predictive model in Phase II of the Harvest for Health program, which will be used to prioritize underutilized crops for commercial development based on their potential for increased public and private investments.

To submit concepts, applicant(s) must submit an online application using FFAR’S Grant Management System.

The portal for submission will be opened on June 12, 2024 and closed August 7, 2024.

Dr. Constance Gewa, Scientific Program Director
cgewa@foundationfar.org

Ms. Rebecca Gyawu, Scientific Program Officer
rgyawu@foundationfar.org