Early one morning in April, a homeless woman sleeping in an abandoned house in Louisville, Kentucky, was jolted awake by a stranger who pulled her bedding over her head and raped her, according to police.

The 29-year-old woman was taken to the University of Louisville Hospital, where nurses examined her for traces of DNA that could identify her assailant.

Such answers in rape cases typically take months, and sometimes more than a year — delays that can be traumatic for victims and diminish the odds of anyone getting prosecuted.

But the alleged attack occurred while the Kentucky State Police laboratory was evaluating a new “rapid DNA” instrument, which is marketed as a way to identify suspected rapists in hours, while victims are still being treated. If the technology works, it could revolutionize the way rapes are investigated in America, where hundreds of thousands of sexual assault kits remain untested and only a third of reported rapes result in an arrest.

As part of the first real-world test of the technology on sex assault cases, hospital nurses in the Louisville case took extra samples from the alleged victim and ran them through rapid DNA equipment. The device developed a DNA profile of a potential suspect in three hours. Within weeks, a man was under arrest.

The case, still pending trial, reflects the power and the potential of rapid DNA testing as it slowly spreads through the criminal justice system. That growth, driven by two competing companies, has unfolded with little government oversight: While the FBI urges caution, and judges have not yet allowed rapid DNA evidence to be presented at trial, the manufacturers have pitched the technology directly to local agencies. Police have used rapid DNA in ways that push the boundaries of standard law enforcement practice, to analyze crime scene evidence and take DNA samples from people suspected of low-level crimes.

That approach has concerned privacy advocates, criminal defense lawyers and some crime laboratory officials, who warn that the unregulated technology remains at risk of mistakes and abuse. These critics caution that the technique is still new and may not deliver the clear results that the companies claim. That is particularly a concern in rape cases, which are complex because they involve mixtures of multiple people’s DNA.

Terri Rosenblatt, who supervises the DNA unit at the Legal Aid Society in New York, said the instruments haven’t been proven reliable on such mixtures. Errors, she said, could lead to false hits, which could make it hard for suspects to overcome a presumption of guilt.

“This blunt-force instrument that is designed to get results fast, especially with sexual assaults, is problematic,” she said.

Proponents of the technology say early tests show it’s working, and critics and government bureaucrats are too slow to embrace rapid DNA, which could one day replace traditional DNA analysis.

“Rapid DNA is a new and disruptive technology,” said James Davis, a former FBI agent who is the chief federal officer for ANDE, one of the two companies marketing rapid DNA to law enforcement. ANDE worked with Kentucky police to test the technology on sexual assault cases, including the one in Louisville. “I think in time, and that doesn’t necessarily have to be a long period,” Davis said, “it will be accepted as the standard.”

ANDE’s competitor, Thermo Fisher Scientific, has not marketed its instruments for use on sexual assault kits, according to crime laboratory officials who have negotiated with the company. Thermo Fisher did not respond to requests for comment.

For American law enforcement, rapid DNA’s appeal is its path around the monthslong backlogs that plague crime laboratories.

The rapid DNA instruments, which cost up to $250,000 and are about the size of a large microwave oven, are designed to be operated by someone without technical expertise. The instruments automate the painstaking, complicated process of analyzing DNA the traditional way.

In the conventional method, trained laboratory analysts extract DNA from a sample, measure how much there is, make copies of it and then run it through an instrument that produces a unique string of numbers. That code is the person’s DNA profile.

Those steps take hours to complete and require the use of instruments across several rooms of a laboratory. Rapid DNA packs it all into one step: A scientist ─ or crime scene technician, or police officer, or anyone, really ─ places the sample on a disposable “chip,” and puts it into the box. In less than two hours, the instrument produces a DNA profile.

The technology performs best in straightforward situations, involving DNA from a single person, law enforcement officials say. Rapid DNA has been used to identify victims of natural disasters, including the 2018 Camp Fire in California. The FBI is also exploring whether it can use the technology to produce DNA fingerprints while getting booked into jail, a test underway at police booking stations in Arizona, California, Florida, Louisiana and Texas.

The booking station model is relatively simple compared to testing evidence from crime scenes, which can include mixtures of DNA left behind by suspects but also by people who had nothing to do with the crime. Identifying a suspect in such samples requires additional processing and interpretation by trained crime laboratory scientists, sometimes with the help of machine-learning algorithms, experts say.

Rapid DNA instruments aren’t considered advanced enough to perform this added work. That is why the FBI doesn’t allow rapid DNA results on crime scene samples — which include sexual assault kits — to be added to its national index of criminal databases, known as Codis. The FBI has “task groups” watching how local and state law enforcement agencies use rapid DNA on crime scene evidence, and how the technology improves.

Even if rapid DNA helps solve a crime, potentially taking an attacker off the streets more quickly, its results alone cannot be used to convict someone. Law enforcement officials and defense attorneys said they do not know of any attempts to get a judge to approve rapid DNA results as acceptable evidence at trial, although ANDE said it is working with prosecutors to try it on a low-level case, such as a burglary. Until then, police must also run samples through traditional DNA analysis to prove the match.

The Orange County District Attorney’s Office, one of the first agencies to use rapid DNA on crime scene evidence, limits the technology to samples with high levels of single-source DNA, such as saliva on cigarette butts or blood on a broken window, said Robert Mestman, a prosecutor in the office’s science and technology unit. It uses an instrument made by Intergenx, now owned by biotechnology giant Thermo Fisher Scientific, the other company besides ANDE that is marketing rapid DNA in the United States. Since 2015, rapid DNA has led to 80 convictions in the county in crimes ranging from murder to burglary.

But not any sexual assaults.

“Sex assault kits are classic mixture scenarios,” Mestman said. “Based on our experience, mixtures are too complex for rapid instruments to handle. So we send those to the crime lab.”

The Arizona Department of Public Safety uses rapid DNA in a similar manner, and it, too, does not apply it to sex assaults, Vincent Figarelli, who runs the agency’s crime laboratory, said.

“The issue with sex assault is that pretty much every case is a mixture,” Figarelli said. He also said that no study showing the accuracy of rapid DNA on sexual assault kits has been published yet in a peer-reviewed scientific journal.

ANDE, a private Colorado-based company, believes that it has found a way to analyze the kits.

While a study by independent laboratories concluded that ANDE’s instrument met FBI standards to produce DNA profiles of arrestees in booking stations, the company said there is no such published research on its instrument’s use on sexual assault kits. But it has made sex assaults a key piece of its marketing ─ and sees the Kentucky project as validation.

Davis said he was confident that ANDE’s technology would one day save government agencies millions of dollars by allowing authorities to mow through the country’s massive accumulation of untested sexual assault kits.

“Rapid DNA without question will reduce the sexual assault kit backlog an hopefullly do away with it,” Davis said.

When ANDE first pitched rapid DNA as an investigative tool for sex assaults in Kentucky, officials were intrigued but dubious.

The state had implemented reforms and spent millions of dollars to reduce its backlog of untested rape kits. But the wait for new kits still lasted about 10 months.

That was untenable: The longer victims wait for news on their cases, the more likely they are to assume no one believes them, advocates say. If the investigation moves quickly, victims remain more engaged ─ which improves the odds of successfully prosecuting their attackers.

ANDE offered a solution. In a meeting at the Kentucky Justice and Public Safety Cabinet in Frankfort in February 2017, ANDE representatives outlined their vision, which had never been tried before in the U.S.: by putting a rapid DNA instrument in a hospital, evidence from sex assault kits could be processed immediately.

The Kentucky officials were skeptical, because, they said, the ANDE representatives didn’t mention how they’d get around the problem of analyzing DNA mixtures. The company also overstated the FBI’s endorsement of its technology, the officials said. They rejected the pitch.

“There was nothing about that project that made any sense whatsoever,” said Regina Wells, the DNA database supervisor for the Kentucky State Police, which runs the state crime laboratory. “They had not thought that through at all. They were just trying to get people to try it.”

It was not the only time ANDE has been accused of overstepping. The company’s attempt to introduce rape-kit testing in Houston backfired this year after the Texas Forensic Science Commission found out that ANDE had arranged with a local hospital to take extra swabs from rape victims to send in the mail for rapid DNA processing. According to the commission, ANDE had acted without the knowledge of local police, prosecutors or the city’s crime lab, the Houston Forensic Science Center. The commission sent ANDE a letter asking that it stop working on crime-scene samples in the state until the company found an accredited government laboratory to work with.

“There’s real value in this. It could save lives,” Peter Stout, CEO of the Houston Forensic Science Center, said. “But how do we do it in a way that protects the rest of the process and doesn’t compromise cases?”

ANDE has said that the Houston Forensic Science Center did know about its activities as they were taking place, contrary to what officials said, and that the company is being held up by laboratory officials who feel threatened by the new technology.

In Kentucky, ANDE did not give up. It admitted its mistakes ─ Davis, ANDE’s chief federal officer, said some sales representatives in early 2017 “were still trying to wrap their head around what we could do” ─ and returned with a stronger pitch in the fall of 2018. The company persuaded authorities there to try out their product.

“We all see ourselves as being the good guys here,” Davis said. “This is all about trying to do the right thing.”

The key to winning Kentucky’s support was ANDE’s effort to solve the mixture problem.

ANDE said it had come up with a process that isolated and extracted sperm cells from sexual assault kits much faster than traditional DNA laboratories. The company had tested the method in its lab but not on sexual assault kits. Kentucky was “the first place where we could validate the science,” Davis recalled.

Wells and her boss, laboratory director Laura Sudkamp, said they weren’t convinced but decided to try out the technology.

“I just wanted to see if I could prove whether it works,” Sudkamp said.

The result was a pilot involving 100 sexual assault cases handled by the University of Louisville Hospital and a few nearby police departments.

The project began in December 2018 and ended in July. In most of the cases, there wasn’t enough DNA evidence, which isn’t unusual. But in 14 cases, the rapid DNA instrument produced a profile of a suspect that could be compared against the state’s criminal database. Four of those comparisons resulted in hits that were later confirmed by traditional DNA — all cases that are pending.

The case that illustrated rapid DNA’s biggest potential — in the investigation of rapes by strangers — unfolded April 27, when the homeless woman called police to the abandoned house in Louisville.

Without a clear description of the attacker, police had no leads.

“We were nowhere with this case,” said Sgt. Tim Stokes of the Metro Louisville Police Department’s sex crimes unit.

Once the sample collected for the rapid DNA instrument reached the state laboratory, it took about three hours to isolate the sperm and develop a profile, officials said. A search of the state criminal database revealed the match with a man with past drug convictions. On May 9, that was relayed to Louisville police, who obtained a previous mugshot of him.

Louisville detectives immediately began looking for the victim. That search took them to a homeless encampment on the edge of town. A detective showed her the mugshot. “That’s the guy,” she said, according to Stokes.

On June 19, police charged Antonio Short, 22, with first degree rape. He has pleaded not guilty and remains jailed on $500,000 bond while awaiting trial, which is scheduled for March 2020. His lawyer did not comment.

Last month, the traditional DNA test on the rape kit confirmed the rapid DNA instrument’s hit, Sudkamp said.

If police had been forced to wait until November to learn of the match, it would have been harder to connect with the victim and pursue her alleged attacker, Stokes said.

“We wouldn’t have had a suspect in custody,” he said.

Eileen Recktenwald, executive director of the Kentucky Association of Sexual Assault Programs, said her organization was “very excited about the possibility” of rapid DNA.

“The quicker the resolution, the quicker victims can get on with their lives,” she said. “When nothing happens, I hear victims say, ‘My life is on pause and I’m looking over my shoulder all the time.’ That’s not a good way to live.”

Sudkamp and Wells describe themselves as cynical and hard to impress, but they now believe in rapid DNA enough to continue working with it. They’ve concluded that while it shouldn’t replace traditional DNA analysis, it’s a viable tool for processing rape kits. They’re sharing their results with other crime laboratories and at conferences, and are looking for funding to expand the use of the technology to other types of crimes.

“This instrument does certain things really well. And that’s what we’re going to focus on,” Wells said. “My skepticism has gone away.”

But Melanie Lowe, general counsel for the Kentucky Department of Public Advocacy, which runs the state’s public defender system, is not convinced. She said she’s concerned about contamination of samples and is distrustful of a method that hasn’t been proven reliable as courtroom evidence — and that defense lawyers haven’t had a chance to challenge.

“DNA is the most valuable piece of evidence juries are looking for. When you attach that cloak of reliability to a new version of it, I’m going to be skeptical until it’s been peer reviewed and tested. Because my clients’ lives are at stake.”

Rosenblatt, of the Legal Aid Society, said she agrees with police and sexual assault victims’ advocates that it’s important to speed up testing. But she sees rapid DNA as an emerging, unproven technology. A better approach, she said, would be to hire more traditional laboratory analysts and expand services for victims.

“Those issues are real concerns, and law enforcement is right to care about them,” Rosenblatt said. “But using not-ready-for-prime-time forensics isn’t the answer.”

ANDE, meanwhile, is working to develop sexual assault projects in other states, which it wouldn’t name.

“The problem here,” Davis said, “is that we’ve got crimes that we could be solving that we’re not.”