Syngenta optimized soybean breeding with four OR tools, avoiding $287M in R&D

Soybean breeding traditionally meant planting huge trial populations and waiting seasons to learn which crosses worked — slow, land-hungry, and expensive. Syngenta's soybean R&D team under Joseph Byrum reframed breeding as a decision problem under its 'Good Growth Through Advanced Analytics' program: use mathematics to pick the best strategy before committing resources.

The team built four operations research tools mapped to pipeline phases. The trait introgression (TI) and breeding project lead (BPL) tools let each project lead compare alternative strategies for transferring favorable traits, iterating until a strategy showed high probability of success at minimal cost and time. The yield trial design (YTD) optimizer and data quality cart (DQC) tools improved the design of field trials and the quality of variety selection and advancement toward commercialization.

The approach raised genetic gain — accuracy, selection intensity, genetic variation, and generation time — without using more land, water, or inputs. In 2015 Syngenta won the Franz Edelman Award, beating finalists including IBM, the Defense Logistics Agency, Ingram Micro, and a US Army/Sandia team — a striking win for an agriculture firm in a field usually dominated by defense and tech.

Score every breeding strategy before planting a row

Breeding has the structure OR is built for: sequential stages, uncertain outcomes, and very costly trials.

Each tool targeted a specific pipeline phase, so gains compounded across design, development, and evaluation.

Project leads could compare strategies in silico instead of paying for failed crosses in the field.

The program measured itself in dollars: $287M cost avoidance and a $1.5B portfolio value made the case undeniable.

Any pipeline with stages, probabilities, and expensive trials is an optimization problem — even one that grows in dirt. Model the decision before spending on the experiment.

The tools produced more than $287 million in cost avoidance for soybean R&D from 2012 to 2016 and improved the probability of delivering a portfolio worth over $1.5 billion. Syngenta then launched a multi-year effort to adapt the soybean tools across all its major crops, and spun off the Syngenta Crop Challenge to crowdsource analytics talent into agriculture.