by Dr. Ryan Boldt, IGS Lead Geneticist
Genetic Evaluation and the Intersection of AI
In the beef cattle industry, we are big on using abbreviations, whether we are talking about EPD, industry groups, or a variety of other things. However, one of the ones I run into the most often is AI. In my world the abbreviation AI can have multiple meanings, and who I am talking to helps me gain context on which AI we are referring to. The first and most obvious is artificial insemination, which can claim the throne of the first implementation of the AI abbreviation in the beef cattle industry. The second use of abbreviation AI is artificial intelligence. This application of AI is starting to become a bigger part of daily life across the globe. Sometimes you are aware that you are interacting with it and other times you may not be aware. For example, going to a specific website like ChatGPT is an obvious interaction with artificial intelligence. However, many companies are using it to form recommendations for content or when online shopping.
Bringing it back to genetic evaluation, artificial insemination has been one of the cornerstones that has allowed breed-wide genetic evaluations to be feasible. The use of sires across different herds allows for a linking of information from each of these herds to one another. As we expand this in the context of IGS, this linking is not just limited to one breed but across different breeds and different continents. In fact, around one-third of animals included in the evaluation have a half-sibling that is from a different database than the one they are registered in. Pooling data not only allows us to create more accurate genetic predictions for those commonly used sires, but also for all the animals and pedigrees that are compared in contemporary groups for every trait we evaluate. The linking of this information is vital for the data structure that makes genetic evaluation work.
The newer “AI” is also something that is becoming increasingly important to genetic evaluation. The IGS genetic evaluation has been using a form of artificial intelligence known as neural network in the evaluation for over five years now. While it is impossible to know, this is likely one of—if not the first—implementation of artificial intelligence in routine beef cattle genetic evaluation. Nowadays, artificial intelligence also helps with the process of developing genetic evaluations. All the tools that I use on a day-to-day basis have artificial intelligence tools included in them to help make writing code for the genetic evaluation easier. While I don’t claim to be an expert on the topic, I believe that we are just on the tip of the iceberg where this technology can be implemented. We have been exploring other ways that this technology can be added to the evaluation and some of those projects are showing some real promise.
While these are the ways that I interact with this technology currently, I also think that there will be new innovations in this space that will help with many aspects of raising cattle in the future. Likely, there will be implementations that help with data collection and phenotype formation in different traits that we may currently be recording or haven’t been able to record previously. I can also see where this technology can be implemented to help with managerial decisions for individual animals. Admittedly, there are a whole host of other possible applications that may prove to be beneficial. Both technologies that use the abbreviation AI have been and will continue to be important to improving many aspects of beef cattle production. .
