b'PARTNER CONTENTSteve TanksleyOperations Research Takes the Guesswork Out of Plant BreedingI ve heard the phrase plant breeding isTraditional Optimizedmore an art than a science. Respect-Reduced Trialing Costsfully, I disagree. Its not easy, but withTrialing Costs (50-90% of R&D Budget)the right algorithms, one can take the mystery out of plant breeding to turn theStage 1: 10,000 new lines Stage 1: 5,000 new linesart into a near-exact science.Think about it like this, when a manu- Stage 2: 1,000 lines Stage 2: 500 linesReduced facturer creates a new car they start with aTime time to new(7-20 concept, then they find all the right piecesyears) Stage 3: 100 lines productsStage 3:and parts, and finally they put it together.20 linesThe mechanic doesnt wander throughStage 4:the shop blindly grabbing parts in the20 lineshope theyll all jive. The designer doesnt leave any of the functions or appearancesNew Products New Productsto chance. Each member of the team gets and gives the new car exactly what itThere are 1.6X1060 (yeah, thats a LOTToday weve worked with 15 different needs to perform. of zeros) possible combinations. It doesntcrops using optimization algorithms find So, what does this example have to domatter how well a scientist knows the 200the best possible solutions for crops, given with plant breeding? Glad you asked. parent linesthe reality is, that persona specific set of parameters. This problem-An emerging field of scienceopera- cannot possibly envision all of the possiblesolving approach creates new genetic lines tions researchuses specialized algo- combinations to find the best options. Ourcalled Optimized Breeding Starts.rithms to not only find good breedingbrains, unfortunately, dont work that way. By creating a more targeted approach, options, but the very best. So, no moreHowever, if we apply operations re- we can take a full step out of the tradi-wandering and blindly grabbing genetics.searchwe can narrow down the 1.6X1060tional breeding process. This saves time By using many data points of information,choices, to a manageable number of tests(obviously) and money for breeders. At the plugging in limitations and parametersof only the best combinations. same time, breeders arent wasting time such as cost and desired outcome, thisNow the art is a science and youre offon crosses that dont produce the optimal field of science has the potential to revolu- to the races to create the best products in aoutcome and instead create only the high-tionize plant breeding. more efficient, and cost-effective way. est quality products.Breeders are limited by the unknown.Plug and play with DNA/genetics.Youre not building a car like my ex-When operations research is applied,Before starting NSIPNature Scienceample earlier, but youre still working with theres no longer a question of what ifImproved PlantsI worked at Cornell de- many parts, pieces and parameters. Dont there is a better genetic combination thatveloping tools around molecular breeding.go in blind when building your genetic we missed? Instead, it propels breedingWhen the human genome was mappedmasterpiece. The crops youre working advancements faster, cheaper and morearound 2000, I started really diving intowith can be optimized just like a caryou accurately. what this could mean for other industries,just have to know how to do it and were Lets take a look at a hypothetical.like crop genetics. here to help with tools only operations Say you have 200 parent lines for wheatIn that research, I quickly learned theresearch can provide.and youre trying to achieve higher yield,biggest challenge to moving the needle onFor more information about how you Fusarium head blight resistance, bacte- genetic advancements, was managing thecan optimize your breeding approaches, rial leaf streak resistance, early vigor andsheer amount of data it created. We startedcheck out nsiplants.com or emailwinter hardiness. NSIP 2006 to address the data challenge. info@nsiplants.com.JUNE 2021SEEDWORLD.COM /23'