Putting the AI in AI

Animal Sciences research team awarded $1 million NIFA grant to improve artificial insemination with artificial intelligence

Peter Sutovsky, professor of animal sciences, and one of his postdoctoral researchers, Lauren Hamilton, were recently awarded a $1.3 million grant by USDA-National Institute of Food and Agriculture (NIFA) to support their work in creating artificial intelligence they hope will not only help farmers and the agriculture industry, but couples battling infertility, too.

“We want to translate this into any area of science where it can be beneficial,” Sutovsky said. “That could be livestock, humans or even endangered species.”

The project itself involves examining tens of thousands of images of sperm samples from 85 genome-sequenced bulls for morphological irregularities and creating artificial intelligence software that can spot these irregularities using image-based flow cytometry. The process will allow farmers to quickly and easily identify bulls whose sperm may not produce a viable pregnancy.

“As we say, we put AI in AI,” Sutovsky said, noting that both artificial intelligence and artificial insemination — which is commonly used in livestock breeding — are referred to by the moniker.

There are many abnormalities that this imaging can detect. A broken tail of a sperm, for example, would result in a sperm that is unable to swim, and therefore unable to fertilize an egg. Double tails and bumps on heads — known as knobbed acrosomes — are also common problems detected by the imaging technology. Sutovsky also hopes to identify sperm-carried DNA mutations related to miscarriages.

“Spontaneous pregnancy loss is one of the biggest concerns for cattle farmers,” he said.

He further explained that if a spontaneous pregnancy loss occurs, the cow must be re-inseminated. That process is not only hard on the cow physically, but it also means that farmers may miss the ideal window for better calving outcomes.

The identification of fertility markers from the male that may contribute to miscarriages could be particularly valuable in treating human infertility, according to Sutovsky and Hamilton.

“We know a lot about female infertility, but a lot of male infertility is unexplained,” Hamilton said.

Sutovsky explained that the ability to identify these fertility markers in sperm could help couples pursuing in vitro fertilization, or IVF — especially when the treatment is resulting in pregnancy, but the pregnancy is lost. This process is both financially and emotionally difficult for many couples.

This project works off the foundational results obtained from a previous USDA-NIFA Dual Purpose Dual Benefit grant that was Hamilton’s focal postdoctoral  project. The USDA-NIFA Dual Purpose Dual Benefit program seeks to fund projects that are poised to benefit advancement for both agriculture and medicine.

However, this is the first USDA-NIFA grant Sutovsky’s lab has received under this grant program focusing on shifting from the population-focused genome studies to the individuals’ phenome, or set of genetically predetermined properties reflected by unique appearances of sperm cells. This represents a shift for the Sutovsky Lab into the field of precision genomics, identifying specific genetic mutations that cause the aberrant sperm defects that are seen through high magnification microscopy and high-throughput flow cytometry in individual animals.  Overall, this project is examining fertility through the lens of phenomics and looking at gene variation on an individual level.

“It was a shot in the dark,” Sutovsky said of the grant application. “I have colleagues who are very successful in genomics in this same round, but I thought it couldn’t hurt to ask.”

Sutovsky credits the help of colleagues at the University of Missouri for making the project possible, especially Dr. Robert Schnabel, associate professor of animal sciences, and his graduate student Jacob Rissman, who have worked to identify the genomic variants being examined within the cohort of 85 bulls used.