MeIt’s 1am on the West Coast of the United States, but the Emerald Cloud Lab just south of San Francisco is still busy. Here, 24 hours a day, 7 days a week, nearly unattended benches spin more than 100 high-end bioscience instruments, conducting experiments for researchers around the world. I am “visiting” via a camera mounted on a chest-high telepresence robot. The 1,400-square-meter (15,000-square-foot) lab is run by Emerald CEO Brian Frezza (Brian Frezza). There are no real scientists anywhere. Several staff members in blue coats quietly follow instructions from the trolley’s screen to ensure that the instrument is loaded with reagents and samples.

A cloud lab means anyone, anywhere can remotely control an experiment using just a web browser. Experiments are programmed through a subscription-based online interface. Software coordinates robots and automated scientific instruments to perform experiments and process data. Friday night is Emerald’s busiest time of day. The scientist plans to spend the weekend relaxing with his family while conducting experiments.

There are still some things robots can’t do, such as lifting giant carboys or opening samples sent in the mail. There are also some devices that cannot be automated. So people in blue coats look like pickers in an Amazon warehouse. In fact, it turns out most of them are former Amazon employees.

Emerald initially hired scientists and lab technicians to help keep the facility running smoothly, but they lacked creativity and had little to do. Pulling out an Amazon employee turned out to be an improvement: “We’re giving him the money he was getting at Amazon to do something much more fulfilling than stuffing toilet paper in a box.” We pay twice as much,” says Frezza. “You’re going full speed ahead with someone’s drug discovery experiment.”

Further south in the San Francisco Bay Area are two cloud labs operated by Strateos. Racks of gleaming life science equipment—incubators, mixers, mass spectrometers, PCR machines—are housed inside large Perspex boxes called workcells. The setup is arguably even more futuristic than the Emerald. Here, reagents and samples are quickly moved to the correct work cell on a high-tech magnetic conveyor belt and gently loaded into place by a dexterous robotic arm. As Marc Siladi, his director of operations executive at Strateos, puts it, the researchers’ experiments are “delocalized.”

Automation in science is nothing new, especially in fields such as molecular biology. Much of the experimental work involves the laborious and repetitive transfer of small amounts of liquid from one vial to another. The disruption caused by the pandemic has led many professional facilities to develop ways to operate their equipment remotely. (For example, the beams of the UK’s powerful diamond light sources, such as particle accelerators that produce ultra-high-energy radiation to probe matter, can now be operated by users from anywhere in the world.) Experimental processes are also new. Not a thing.

But Emerald and Strateos are different. These are the worlds that theoretically allow anyone with a laptop and a credit card to “pay and play” for the entire reagent inventory and suite of instruments available at world-class research facilities. It is the first research institute. The appeal of this approach became apparent during the pandemic when many researchers were unable to visit their labs in person. Cloud Labs founder says this is the future of Life Science.

The most obvious advantage is productivity. Researchers can run multiple experiments at once and queue them up all night or while doing other things. “Our professional his user, in a conventional lab he does the work of 10 scientists,” he says Frezza. “They will put out ridiculous numbers.”

No time spent setting up and tearing down equipment, cleaning up, maintaining and repairing equipment, or replenishing inventory. Arctoris, a remote drug discovery lab in Oxfordshire, says its platform has allowed him to complete a pharma project in 24 hours that would have taken him at least a week in a traditional environment. Instead of spending hours pipetting every day, researchers can spend more time thinking, reading, and analyzing results with colleagues.

Scientists at Carnegie Mellon University in Pittsburgh have been very impressed with what the Emerald Cloud Lab can do for staff and students. A researcher was able to replicate his long-standing Ph.D. experiments in just a few weeks. for them.

Rebecca Doerge, Dean of the School of Science at Carnegie Mellon University, said that a year of access to a cloud lab is often cheaper than the price of a single piece of high-end lab equipment, so this model is transformative. said to have the potential to bring about “I’m not just interested in changing science at Carnegie Mellon University. I’m interested in changing scientific processes around the world,” she says of the new facility in Pittsburgh. “I have colleagues in underresourced places who can’t do the science they can, just because they don’t have enough money. With internet connectivity and cloud access to her lab, this is a game changer.”

Doerge, a statistician turned science manager, is also excited about removing variability and human error from experimental work. There are no scientists on the new 1,500-square-meter (16,000-square-foot) site, and he has only six engineers to help the place run 24 hours a day. “People still go to wet labs and stand there and make mistakes. The thing is, once you learn it, you don’t have to stand there and repeat it over and over again.”

Scientists such as Doerge believe the precision of remotely controlled labs can help solve what has come to be known as the “reproducibility crisis” in science. This is a disturbing revelation that if different groups of scientists performed the same study, they wouldn’t be able to replicate the mountain of published research results. method exactly. When an experiment is plugged into a browser and run on a robot, researchers need to translate the exact details of each step into clear code. For example, what was once described in scientific papers as “mixing a sample” becomes detailed computer instructions for a particular machine to mix at a particular number of revolutions per minute for a particular amount of time. Other factors that can affect the results, such as the ambient temperature at the time, are captured in the metadata.

Doerge has encouraged more and more research and even teaching positions at Carnegie Mellon University to move to remote labs, but not all of her colleagues have been supportive. Many scientists work with their colleagues on benches and believe that the sights and sounds of an experiment help generate exciting ideas and happy coincidences. Others are concerned about the quality of data generated in non-existent labs. faculty,” she says Doerge. “It’s certainly a change of mindset.”

Some experts believe that making access to sophisticated labs this easy poses a potential biosecurity or bioterrorism threat. In theory, even small groups and individuals with no research experience can use Cloud His Lab to initiate complex biological experiments. “The lab says it only works with trusted partners, but of course we are very keen to develop the market,” said Kings, a biological risk and biosecurity expert at his College London. said Dr. Philippa Renzos. “You have to remember that most people are from good places, but there are some people out there who are pretty crazy. increase.”

Cloud Labs says it has a system to review all planned experiments and flag or deny those deemed illegal or dangerous. Additionally, they argue that fully digitalizing everything that happens in a lab will make it easier to record and monitor what people are actually doing than in traditional labs. .

Paul Freemont, co-founder of the UK Innovation and Knowledge Center for Synthetic Biology, is developing several highly automated labs, including a robotic platform capable of conducting more than 1,000 Covid tests a day early in the pandemic. has supported He isn’t sure that remotely operated laboratories are still “mature” enough for scientists to set up their own automated equipment to recreate what’s available. “I like the concept, and I think this is the way science should go. It works if you have all the necessary protocols and workflows that a biologist might need, but right now, biologists I don’t think it’s available at the level of complexity and detail that is needed.”

Fremont is also concerned that scientists don’t really understand or get involved with the software and hardware that generate the data. “We need to get the next generation of scientists to understand how to build all this infrastructure themselves and how to operate it. I don’t think it’s very healthy for a large private company to monopolize that understanding.”

Despite these concerns, the demand for cloud science is growing. Emerald is expanding its production capacity primarily to meet demand from pharmaceutical companies and biotech start-ups. Strateos is working with the US research institute Darpa to study in detail how the facility can improve the reproducibility and efficiency of previous experiments, and the company also licenses the software to other institutions are allowing facilities to be converted.

In the future, cloud labs may even decide to do their own experiments. As Google’s DeepMind platform recently proved, machine learning tools can swallow decades of data and spew out answers to questions that would take scientists years to solve with physical investigation. became. Pharmaceutical companies are increasingly using these tools to simulate molecular interactions in their search for new drugs. Data generated through cloud labs translating biology into information technology only makes these tools more powerful. Combining all these technologies could one day lead to systems where theories can be developed and physically tested without human input.

Already, some advanced users of the Emerald Cloud Lab have developed algorithms to adjust the parameters or directions of their next experiment based on their own data analysis. “It’s kind of wild and very futuristic,” he says Frezza.

All this means that scientists are the latest profession to ask what the move to automation and AI means for the future. Is it possible that one day more traditional research scientists will lose their jobs? After all, you will always need someone to prioritize the questions that need answers and develop new ways of answering them. But the days of sitting on benches in lab coats and gloves by the fire of a Bunsen burner may soon be a thing of the past. The era of robotics researchers is coming.