Minnesota began analyzing more samples from COVID-19 patients this fall than almost any state, intensifying its search for new coronavirus variants that could worsen illness or undermine vaccination.

Roughly 8% of Minnesota's 920,000 coronavirus infections have been sequenced to identify their genetic fingerprints, but public and private labs are now analyzing 20% to prevent any troubling variants from surging undetected.

At 2,000 genetic sequences per week, Minnesota is likely to detect a significant variant even if it exists in 0.1% of samples, said Sara Vetter, director of the state public health laboratory. "If something is circulating around in 1 in 1,000 [samples], we will find it."

The tighter safety net helps explain why Minnesota was the second state to identify an infection involving the omicron variant a week after it was discovered in South Africa. The infection involved a Hennepin County man who experienced mild COVID-19 symptoms Nov. 22 after attending a convention in New York.

Minnesota has submitted the third-most genomic sequencing results of coronavirus samples over the past 90 days to the GISAID open-access database, which allows for broader research for trends among the results. Colorado also was among the first five states to find omicron infections, and it ranked second.

"It could be coincidence ... but I don't think so," said Kenny Beckman, director of the University of Minnesota Genomics Center, which is part of the state's sequencing group.

Genomic sequencing this summer showed the shocking rate at which the fast-spreading alpha variant was swept aside by the faster-spreading delta variant. Surveillance was mundane by comparison this fall when delta turned up in 99% of samples.

"In three weeks, we went from no delta to almost all delta," Beckman said

Omicron was labeled a "variant of concern" after it was discovered Nov. 24 in South Africa because it appeared to spread even faster than delta. It also had genetic features suggesting it might evade immunity from previous infection or vaccination.

The increase in sequencing this fall put Minnesota in position to quickly search its breadth of samples to find out if omicron was already in the state.

Minnesota's sequencing partnership includes the state public health lab and the U's Genomics Center, but also Infinity BiologiX, a New Jersey firm that processes samples from the state's saliva-testing centers. Fortunately, they had a good idea where to start, because some variants, including alpha and omicron, react unexpectedly to the tests, Beckman said.

Standard PCR diagnostic tests seek multiple genetic targets to identify a coronavirus, and the goal is to find them all and leave little doubt of the results, Beckman said. With omicron, some tests fail to identify the S-gene target even though they find other targets and confirm the infections.

"You actually have a clue, which is quite nice, that's coming from some of the commercial tests failing" in a way that produces helpful information, said Beckman, whose center also sequences samples from Arkansas, Missouri and South Dakota.

Eight samples were identified, including from the Hennepin County man whose infection proved that omicron was in the U.S. before it was announced in South Africa. Sequencing of at least four other samples did not find the omicron variant.

Omicron isn't the first variant to raise concerns about a mutation that could allow the coronavirus to evade immunity. The beta and gamma variants found earlier this year in Brazil and South Africa, respectively, raised similar concerns and were eventually detected in Minnesota. But delta proved the stronger variant, and sequencing in Minnesota found only 262 infections involving beta and 552 involving gamma.

Whether omicron can surpass delta as the dominant strain is unclear, but sequencing should provide answers in four weeks, said Matthew Binnicker, vice chair of practice for Mayo Clinic's Department of Laboratory Medicine and Pathology.

"Delta showed us that it was the king of hill. It could outcompete other variants and became 99% of the virus in the country," he said. "We'll have to see ... if omicron is better at being transmitted than delta."

Genomic sequencing has long been used by epidemiologists in Minnesota, particularly in the search for sources of foodborne diseases. During the pandemic, sequencing identified how participants of the 2020 Sturgis Motorcycle Rally in South Dakota brought the virus back to Minnesota.

Sequencing also allowed for a detailed map this spring that linked coronavirus infections among multiple schools and sports teams in Carver County and surrounding areas of the south metro.

No variants labeled as concerning have been discovered first in Minnesota. Beckman said such variants are typically found where they arise. A variant of interest was identified earlier this year in California, but limited genomic sequencing early in the pandemic reduced the discovery of other variants in the U.S., he said.

"I think the United States is now sequencing at a frequency such that if there are variants of concern here, they will be detected," he said.

Even sequencing 20% of specimens, state health officials said they have to be strategic in which ones are selected in order to increase the chances of finding variants of concern. A random sampling of specimens is selected from across Minnesota to provide geographic monitoring, but epidemiologists also can recommend sequencing of suspicious clusters — such as from communities with surging transmission rates. Sequencing also has tried to identify any variants more likely to cause hospitalization or breakthrough infection in fully vaccinated people.

Matching genomic sequences "help connect the dots," Vetter said. "So then if we see something in a school, if we see something in a geographic area, we can see how [the virus] got introduced and spread around the community."