iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Computomics CEO Sebastian Schultheiss says the company’s xSeedScore platform uses genetic, field-trial, environmental, and farm-management data to predict crop performance and help breeders decide what to test next. In this interview, he explains how the approach accounts for growing conditions, how predictions are evaluated, and what the company says about protecting customer data. The interview describes the company’s methods; it does not provide independent performance or security assessments.
What sets Computomics’ approach apart from other companies in the plant breeding industry?
Schultheiss contrasts conventional linear mixed models, which estimate how genetic markers contribute to traits, with nonlinear machine learning intended to detect more complex relationships among markers and interactions between genetics and the environment. Drought and temperature, for example, may affect how a plant’s genetic traits show up in a field.
In Schultheiss’s description, xSeedScore treats water availability, temperature, soil conditions, and the growing season as important inputs—not background variation to average away. The aim is to predict how breeding lines might perform in locations or climates that a breeding program has not yet tested, helping breeders choose field trials that are likely to be informative.
The interview gives no head-to-head accuracy study or quantified improvement over other methods, so it supports an explanation of Computomics’ approach, not a ranking of vendors or proof of better results.
#1 Best Overall
What data does Computomics use to predict how a crop will perform?
The interview describes three broad data groups. Their value comes partly from interpreting them together: low yield, for instance, could reflect drought at a trial site rather than a plant’s genetic potential.
| Data group | Examples described in the interview | What it helps assess |
|---|---|---|
| Genotype | Genetic-marker data or whole-genome sequencing, plus pedigree information about ancestry and relationships among breeding lines | Which inherited genetic differences may be associated with traits or performance |
| Phenotype | Field-trial observations such as yield, disease resistance, and crop quality, collected across years and locations | How plants actually performed under measured trial conditions |
| Environment and management | Weather patterns, soil characteristics, water availability, and farming practices at trial sites and during growing seasons | What conditions may have shaped the observed performance |
Environmental records help put trial results in context. If the relevant weather or other site conditions were never recorded, however, the missing facts cannot be reconstructed after the growing season.
Rank #2
How do you know when a model’s prediction is reliable enough to guide a growing decision?
Schultheiss says Computomics tests predictions by withholding whole growing locations or years, using approaches he identifies as leave-one-environment-out and leave-one-year-out validation. This is intended to test whether a model can work in a setting unlike the ones it learned from. By contrast, withholding a random selection of plants can leave related plants or similar growing conditions in the training data, making the test less representative of a future decision.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →He also describes checking calibration—whether stated confidence corresponds to how often predictions prove correct—and assessing uncertainty. A breeder choosing the strongest candidates may not need a model to rank every plant perfectly. Candidates with uncertain predictions can be tested in the field rather than automatically discarded.
Rank #3
The interview reports no numeric validation results, so it does not establish a particular accuracy level or guarantee that a prediction will hold in a new growing season.
How does Computomics protect proprietary plant data while still generating useful predictions?
Schultheiss says customer data is processed on Computomics’ own infrastructure under data-processing agreements, separated by customer, and used to train models for individual customers or specific breeding programs. He says the company does not pool genetic material across clients.
He gives a practical reason for that separation: combining customers’ breeding material might lead to a suggested cross involving a line that one customer cannot access or has no legal right to use. He says public reference genomes, environmental data, and Computomics’ own modeling techniques can contribute to improvements without exposing a customer’s proprietary genetics.
Recommended Free Tools
These are Schultheiss’s descriptions of company practice, not an independent review of infrastructure, contracts, or security controls.
Best Value
What is the biggest misconception about using machine learning models in plant breeding?
Schultheiss addresses three misconceptions about how genomic prediction fits into a breeding program:
- It eliminates field trials. It does not. Field-grown, measured plants provide evidence for the models, and trials remain necessary to test candidates. As Schultheiss puts it: “In reality, our predictions depend on data from plants that breeders have grown and measured.”
- It changes a plant’s DNA. Genomic prediction estimates which existing genetic variation may be promising; it does not alter DNA. Breeders still decide which parents to cross.
- All historical data must be perfectly organized before a program can begin. Some missing measurements, inconsistent trait definitions, and renamed sites may be manageable. But environmental facts that were never recorded—such as a site’s past-season weather—cannot be recovered later.
Schultheiss made these comments in David Kertai’s Center for Data Innovation interview, published October 2, 2026: Center for Data Innovation.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches
