Bob Seger and the Silver Bullet Band released the song “Like a Rock” in 1986. The song soon became incredibly popular, resonating with the American ideals of personal resolve, strength, determination and a rugged individualism.
Chevrolet decided to harness this wave of popularity and incorporate the song in the marketing campaign for their Chevy truck ads as of 1991. You can listen to their ads with Bob Seger singing in the background. Chevy trucks – more horsepower, more reliable, more dependable and longer lasting than all other trucks. Why? Because they took the time to build their trucks “like a rock”.
This “built like a rock” concept can also apply to plant breeding software. How? Because a simple database architecture will not deliver the results for a serious plant breeding program. The database, which is the foundation of everything else, really must be “built like a rock”. The database must be highly relational so that it links all the data in many ways that really empower the breeder. Most software developers for such software do not take this seriously enough, starting with a simple database then spending more time on the interface, appearances, and visualization tools which can be spellbinding to those who don’t take the time to look deeper.
In Chevy truck terms, if the chassis is weak and there are only four cylinders, a shiny cab and fancy controls will never be able to compensate. It looks great, but will never deliver the results. Or if you are not a truck “enthusiast”, what would you think of a BMW that happened to have the smallest Ford Fiesta engine? The “performance gap” would be noticeable.
Noticeable performance costs
I have visited a number of seed companies that have licensed other commercial plant breeding software and they emphasize that the software is not really relational, despite claims otherwise. Certain results cannot be delivered, some operations are difficult or take hours and opportunities are lost. In some cases, company revenues have been declining or remain flat because there have not been enough superior hybrids or varieties to stay long enough in the market. One company showed me a detailed Power Point presentation to make the point, a real concern for the company’s management. The “performance gap” is noticeable, and it costs the company.
Let’s test it
So how do you know if your plant breeding software is truly relational? Let’s start with a few simple tests:
Can you change the name of a genotype in one place in the database, and the new name appears instantly in even hundreds of experiment or nurseries? Or how about a trait – can you change the name of a trait in one place, and then everywhere it appears in the entire database all experiments and nurseries open up automatically with the new trait name? If you cannot, there will certainly be practical implications that you cannot avoid and will cost you time and effort.
Can you load all the data you ever took for a given trait, or a given genotype, or an entire year or location with just a few clicks? If not, some basic management operations will take more time every time, if it can be done at all.
Linked, relational data
Now look deeper at the ability to link or relate data. Can you easily link, with a few clicks, genotypic or marker data for an entry in an experiment or nursery (or groups thereof) with the plot data, as well as experiment or location attributes and related seed lot traits? This allows more possibilities for data calculations, monitoring response to selection and meta-analyses. Or how about linking data from past generations and one or both parents for a population in a breeding nursery? In all my training courses, breeders have really appreciated this capability and see real value. Opportunities are lost if you cannot link and mine all your data for more powerful analyses or breeding and yield trials decisions. In short: maximum use of all your data for maximum results.
All the above capabilities can help minimize or eliminate potential legal costs. What if a gene is licensed from another organization, or a certain genotype as a parent in crosses? The ability to trace every cross or population where the gene or genotype is used, every plot where planted, every bit of seed, and all done reliably with a few clicks, makes external audits and royalty calculations really easy. Some of our client seed companies used to complete such audits over a few weeks, with nervous breeders and management. Now with our software, the audit takes only a few hours and the staff in the companies are quite relaxed. Real value, not many people look forward to paying lawyers when they could have avoided it. A simpler database architecture will not “deliver the goods”, not even close. The interface might look great, but there will be a performance gap.
I speak from personal experience when I say I’m very grateful for a solid, well-built vehicle. I do confess, though, that I do not drive a truck or a Chevrolet. My car was involved in 2 bad accidents within 2 years due to the fault of others. Without a solid-built vehicle I could have experienced terrible tragedy.
So is your plant breeding software “built like a rock”? If not, you can be sure that it will cost you in the long run.
Categories: Plant Breeding