Christine Baes gave a presentation titled “The genomics of fertility” during the reproduction session at the 2026 Western Canadian Dairy Seminar (WCDS) in Red Deer, Alberta, on March 11. Baes – a professor and Canada research chair in livestock genomics in the department of animal biosciences, and associate dean of external relations of the Ontario Agricultural College at the University of Guelph – summarized recent findings and current research in the dairy cattle fertility space.
“When it comes to genetic selection for fertility, it’s not easy,” Baes said. “It is quite slow compared to more straightforward traits like milk production or fat or protein content. This is a complex thing.”
Measuring fertility
Baes opened her session with a quick discussion on traits used to measure fertility: age at first service, nonreturn rate, calving to first service. Baes noted that while fertility is complex, the traits currently used to describe fertility are not. She showed a figure illustrating trends of traits used to breed for fertility and shared a few ideas gleaned from this information.
“The first idea is that our trait definitions are a little too far from the biology of the cow,” she said. “That’s pretty clear. Also, management of reproductive performance might make bad cows look good. Inbreeding depression could have an effect on cow fertility.”
Baes shared work she and her colleagues did on comparing genetic and phenotypic correlations, noting the different ways to measure inbreeding – by pedigree and by genomic testing.
“Looking at the low-inbred cows, they had an average age of first service at 450 days,” Baes said. “The animals with a higher inbreeding level needed six days longer before they could be bred the first time.”
Timed artificial insemination
Baes highlighted research by Gerson Oliviera and Colin Lynch looking at timed artificial insemination (A.I.) records. Oliviera simulated data over 20 generations of cattle, comparing the genetic potential that could be achieved without using timed A.I. programs versus using timed A.I. programs. This simulation showed a stark contrast between the two breeding methods.
“That’s what we’re losing by not differentiating between the animals that are synchronized [from a genetic standpoint] and those that are not,” Baes said.
The research team then dove into this study using real data from DairyComp and analyzed the results.
“You might have a completely different ranking if you’re looking at animals that you have sorted and ranked according to whether or not they were bred with timed A.I. or whether they were just bred using heat detection,” Baes said. She noted that failing to record how a cow came into heat or failing to note that in genetic evaluations is altering genetic rankings.
“Performance with timed A.I. is not the same as performance based on heat detection,” she said. “If we think about these ‘fertility traits,’ we’re not even breeding for these traits – we’re breeding for how well our cows respond to timed A.I. programs. If that’s what we want, it’s OK, but recording those different management practices will help us do a better job selecting truly fertile animals.”
Traits based on reproductive physiology
Baes then discussed research being done to develop new “closer to biology” fertility traits, like analyzing size and position score of reproductive tracts.
“Animals with a small and high reproductive tract had higher pregnancies per artificial insemination by size and position score than those animals with a really large and deep reproductive tract,” Baes said.
Additional studies looked at the heritability of size and position score of reproductive tracts – showing 10% heritability. Anogenital distance – distance between anus and clitoris of a cow – studies showed a shorter distance had more favorable correlations with positive reproduction outcomes.
“We're interested in the genetic part of this and whether or not we can actually estimate what the heritability is of this trait because it’s easy to measure,” she said. “We saw heritabilities of 0.4 – that’s comparable to milk production.” She noted that more studies need to be completed before suggesting the industry selects for this trait.
Baes discussed another influence on fertility – heat stress. She highlighted a study that analyzed at what temperature-humidity index (THI) threshold in different parities increases the risk of embryo loss. This study also showed that animals on timed A.I. programs are affected more by heat stress than those inseminated with heat detection – a difference in THI of 64 versus 73.
“You may think I’m on a crusade against timed A.I.,” Baes said. “I’m not. We just have to understand this better.” She illustrated her point by dividing cows into two groups – those who return to cyclicity easily and those who have fertility disorders – like difficult calvings and metabolic disorders. The cows that are healthy and return to cyclicity quickly will easily fall within voluntary waiting period parameters and be bred without issue.
“This second group are the problem cows,” she said. “I would argue as a geneticist, you have to be ice cold. These cows don’t have any place in your herd. These are the ones you need to cull, not used timed A.I. on.” Baes also noted the influence of bulls with the worst breeding values for fertility disorders, citing data that show a lower performance in daughters of these sires.
A call for accurate recording
Baes then called for attendees to participate in the Dairy Farmers of Canada research project – looking for and reporting genetic anomalies or conditions seen on farms.
“A previous speaker today commended the Canadian dairy industry for having such fantastic data recording, and it’s really true,” she said. “Canada is a leader in dairy genetics and genomics, and if we want to stay there, we better stay hungry and disciplined in recordkeeping.”
“My summary overall is that genomic selection can indeed boost fertility traits,” Baes said. “We’re working on novel phenotypes that allow improved selection for these important traits. We have to keep our eye on the prize of being data-driven and [we] need accurate data recording; I know that’s hard, and it costs money and time, but it’s incredibly worthwhile.”
Q&A session
Baes thanked her collaborators and WCDS for inviting her to speak, then moved to the question-and-answer portion of her presentation.
Q: How are you adapting the Canadian index to improve the fertility of cows? What traits should producers be focusing on when selecting sires within their program?
BAES: We’re focused on continuous improvement. The Canadian index doesn’t change overnight; it’s built on a lot of research and validation, but when updates are made, they’re meaningful and well supported. The goal is to keep refining the traits we include so they better reflect real biological performance and on-farm outcomes.
In terms of what producers should focus on, the current fertility traits are solid. They’ve delivered measurable progress. For example, we’ve seen improvements like shorter calving intervals, so selecting for them will not hurt you. That said, they’re not perfect, and there’s still room to improve how well they capture fertility.
Practically, producers should continue using the fertility traits available today, but be thoughtful about how much weight they place on them in their overall selection strategy.
It’s also important to recognize that genetics is only part of the equation. Management plays a major role in fertility. If management is not optimized – and I am thinking of things like heat detection, nutrition, consistency and overall cow health – then genetic progress will be limited.
For producers already operating at a high level, that’s where genetics can really help fine-tune performance. But for most herds, getting the basics right will have the biggest impact, with genetics providing an additional layer of improvement.
Q: You commented on maybe changing how events are recorded in management software. How do you envision this recording occurring?
BAES: Honestly, the easiest fix is at the software level. If DairyComp or other systems tightened up their drop-down menus and forced more standardized recording, most of this problem would go away.
But I get it. On farm, you’re busy. You’re breeding cows, it’s cold, things are moving fast and recording data is the last thing you want to spend time on. So whatever we ask people to do has to be quick and simple, or it won’t happen.
That said, we have to be honest: A lot of the data being recorded right now isn’t very useful because it’s inconsistent or unclear. If one person writes “treated,” another writes “shot” and someone else writes nothing, we can’t do much with that.
The ask is pretty simple: Record what actually happened, and make it explicit. If you bred a cow, record a breeding. If you gave a treatment, record the treatment. Not shorthand, not something that only makes sense to you or your dad, something that anyone can interpret later.
This isn’t about adding work. It’s about making the data you’re already collecting actually usable.
Q: How do we keep genetic defects and inbreeding from becoming an emerging problem?
BAES: This is something that's very close to my heart.
This isn’t a new problem, and we already know how to manage it. We just actually have to do it.
First, we need to record and report defects properly. If calves are born dead or with abnormalities and it never gets written down or shared, we’re flying blind. You can’t manage what you don’t track.
Second, we need to use the tools we already have. Genomic testing, mating programs and inbreeding metrics are there for a reason. Ignoring them or pushing too hard on a narrow set of sires is how you create problems.
And let’s be clear, every animal carries some deleterious mutations. That’s normal biology. The issue is not that defects exist, it’s how we manage them. When those defects start showing up more frequently, that’s on us.
There’s also a bigger piece here. When we produce animals with obvious defects, that’s not just a technical issue, it’s an ethical one, and it matters for public trust. People don’t accept that anymore, and frankly, they shouldn’t.
So the path forward is straightforward: Record it, report it and manage it at a population level. That means national systems, shared data and responsible selection decisions. If we do that, this stays under control. If we don’t, it becomes a real problem.
Q: What specific traits are contributing to fertility? What about the contribution of other indirect traits like immunity or liver metabolism or overall energy metabolism?
BAES: If you've got an idea for a good trait that you feel is directly involved with fertility, please let us know. But the short answer is: Fertility isn’t driven by one trait; it’s a system.
The core fertility traits we use now are a good foundation, and they’re improving. But we know they don’t capture everything, which is why we’re actively looking at new traits.
If there’s a trait that is biologically relevant, measurable at scale and has genetic variation, we can use it. That’s really the bar. The challenge isn’t ideas, it’s finding traits that are practical to measure across the industry.
That’s also why indirect traits matter. We don’t always measure fertility perfectly, but we measure a lot of things that influence it. Traits related to health, metabolism and resilience are already being picked up indirectly through the index because of genetic correlations.
And this is the key point: We’re no longer selecting for traits in isolation. The Canadian index is built on a balanced breeding goal with a large number of traits – over 100 recorded per cow – all integrated into one system. So fertility is being improved both directly and indirectly.
At this point, it’s not something you can calculate in your head. The math behind these indices is complex, combining direct fertility traits with all of these correlated responses.
So yes, indirect traits like immunity and metabolism absolutely contribute. The question isn’t whether they matter, it’s whether we can measure them well enough to include them more explicitly going forward.
