The Brewing Crisis in Broiler Breeder Fertility
Inside our lab’s effort to map a slow‑moving fertility crisis, using multi‑omics, deep learning, and mathematical modeling to re‑engineer how we can keep male broiler breeders sustainable.
Breeder fertility is key to food security
Each year in the U.S., around 12 billion hatching eggs are set to produce roughly 9 billion broiler chickens a quiet, continuous process that underpins affordable protein for millions. Yet behind that apparent stability, our analyses show a persistent erosion in the reproductive engine that makes it all possible: broiler breeder fertility. Using a decade of national production data, our group projected that hatchability could fall from about 75% today to near 60% by 2050 if current trends continue, with profound implications for food security and production costs. Texas A&M AgriLife feature
In the Athrey Lab, we are treating this as a systems challenge that spans sperm cell morphology, flock management, genetic selection, and the physiology of aging males. Our work brings together multi‑omics, high‑throughput phenotyping, and computational modeling to understand how every missed fertilization event fits into a much larger biological and economic story. PubMed
Why broiler breeder fertility matters
- Broiler breeders are the parents of all commercial broilers; every chick starts as a breeder egg. PubMed
- Hatchability is the share of fertile eggs that result in live chicks; small percentage changes translate into millions of birds gained or lost. Texas A&M AgriLife feature
- Our Broiler Breeder Performance Index (BBPI) integrates hatchability, chick livability, and efficiency to track the long‑term health of this system. PubMed
“Fertility is the bedrock of modern broiler production,” as we wrote in our Poultry Science analysis; once that foundation shifts, everything built on it must adapt.
Reading the warning signs in national data
In our paper “How concerned should we be about broiler breeder fertility declines?”, Cash, Witherspoon, and Athrey use USDA National Agricultural Statistics Service data from 2013–2022 to quantify a trend the industry had long suspected. Our Markov Chain Monte Carlo modeling revealed steady declines in hatchability, chick livability, and production efficiency over that period, even as the number of eggs set and broilers raised climbed to meet rising demand. our published analysis
To translate these trajectories into a single, interpretable metric, we developed the Broiler Breeder Performance Index (BBPI), testing both Gaussian and Cauchy‑based models to evaluate predictive performance under uncertainty. BBPI projections dipping below baseline levels signal that the system is moving away from its historical “comfort zone,” reinforcing the need for targeted interventions in flock management, nutrition, and selection. industry coverage
From data curves to living cells – phenotyping sperm over time
National‑scale forecasts tell us that fertility is changing; our phenotyping work asks why. In his M.S. thesis, “Identifying Spermatozoa Features for Fertility Scoring of Broiler Breeders,” Kolton Witherspoon combined detailed morphological analyses of chicken sperm with advanced analytical methods to pinpoint cell‑level traits linked to fertility outcomes in individual males. These efforts move the field from subjective impressions of “good” semen quality to objective, data‑driven scoring systems that can guide breeder selection and monitoring. Texas A&M AgriLife feature
Building on this, our team has begun tracking sperm morphological features not just as single snapshots, but across time within the same males. By following how traits shift across the reproductive lifespan of a rooster, Triumph Okuku’s work uses temporal sperm morphology profiles to estimate peak reproductive performance windows, providing a new way to schedule use, spiking decisions, and replacement strategies. the paper
Deep learning and AI – letting the data speak
Classical statistics struggle when hundreds of correlated traits, images, and time points collide; this is where our deep learning work steps in. At the Poultry Science Association 2025 meeting, Witherspoon presented our application of deep learning models to dissect which combinations of sperm traits, management variables, and temporal patterns best predict real‑world fertility outcomes in broiler breeders. These models learn subtle, non‑linear relationships that are hard to detect with conventional approaches, turning years of field and lab data into actionable predictions about male performance. the paper
The same computational mindset extends across the lab, where simulation modeling, AI‑assisted image analysis, and predictive indices help us test “what‑if” scenarios long before a hatchery implements them. Whether forecasting the impact of changing male‑to‑female ratios or evaluating new scoring schemes, our goal is to provide decision tools that are both biologically grounded and operationally realistic. open-access article
Multi‑omics – connecting genes, physiology, and performance
Fertility is not just about what we see under the microscope; it is rooted in genomic architecture, regulatory networks, and metabolism. Across the lab, we apply whole‑genome sequencing, transcriptomics, and metabolomics to dissect complex traits in birds, and these same multi‑omics strategies are being deployed to understand male broiler breeder fertility. By integrating omics layers with phenotypes such as sperm morphology, testis condition, and age‑related changes, we aim to identify biomarkers that can be used in selection schemes or early‑warning diagnostics. the paper
This comparative omics perspective sits within a broader research program that spans gut–brain–immune signaling in chickens, mosquito vector biology, and insects as vectors of antimicrobial resistance, giving us a unique vantage point on how networks of genes, microbes, and environments shape animal performance. Lessons from other systems—such as how stress pathways alter gut–brain communication in chickens or feeding behavior in malaria vectors—inform our thinking about how environment and management might intersect with male fertility. the paper
Building a team around a grand challenge
Solving a problem of this scale requires people who are fluent in both biology and data. When Cara Cash joined the lab in 2023, she brought a background in poultry science and a growing interest in data science and smart agriculture, sharpened during a summer program at NC State. Her work on simulation modeling and analytical approaches to broiler breeder fertility and hatchability helped crystallize the need for rigorous, systems‑level thinking about this problem. open-access article
Today, our broiler breeder fertility efforts weave together graduate projects on statistical forecasting, deep learning, and temporal phenotyping, alongside industry‑facing talks and interviews that help translate findings back to stakeholders. From invited conferences and breeder roundtables to coverage in industry magazines, we are building not just datasets and models, but a shared language for talking about fertility risk, resilience, and innovation in the broiler breeder sector. our research program
“Our goal is to treat every data point from every modality, from molecules to the flock level, as a signal in a larger system that we can measure, model, and ultimately change.”
G. Athrey
As demand for poultry meat continues to rise, the industry faces a choice: react to declining fertility with incremental adjustments, or invest in a new generation of tools that make male reproductive performance visible, quantifiable, and improvable. Our lab is committed to the latter—combining multi‑omics, high‑resolution phenotyping, AI, and mathematical modeling to rebuild broiler breeder fertility on a more resilient foundation. Texas A&M AgriLife feature