A Danish startup that wants to help automate lab experiments that require visual inspection by research and development teams has raised $20 million in a Series A round to expand its technology in the U.S.
Founded in Copenhagen in 2018, Reshape is a robotic imaging company packed with software and AI models to help scientists track visual changes such as color and cell growth rates from Petri dishes and similar plate formats. developed the system. The machine has a built-in incubation function that can be set to a specific temperature and records the corresponding data, allowing experiments to be easily repeated.
The advantage is that these experiments can be run 24/7 without direct supervision, freeing up technicians for other important tasks.
Reshape machine in action Image credit: Reshape
“Deciphering nature”
The concept of “deciphering nature” is at the heart of what Reshape is trying to achieve, and builds on a broader trend in which the lines between the natural and man-made worlds are blurring. Opportunities like these are not lost in Silicon Valley, as evidenced by the countless amounts of money poured into technologies that seek to “engineer” biology.
“Biology as a whole is moving from a science to an engineering field. One of the biggest things we want to do is to understand the very 'intangibles,' how things grow and I think it's about realizing how it works. It’s easy to explain,” Reshape CEO Karl Emil Gron told TechCrunch. “Ideally, we want to figure out how to create a layer of translation between what's happening in the real world and what's happening in our DNA.”
Reshape's origins began when Grøn, who has an engineering background herself, started dating a man who worked in the biotech industry, who gave her insight into the amount of manual labor involved in lab experiments.
“We always thought biotechnology was automated on a large scale, but every eight hours of every day for five months straight, she had to go into the lab and take pictures of petri dishes.” said Gron. “When you're in the tech industry, that just seemed crazy.”
After speaking with a number of biotech companies on the ground in Copenhagen, Gron realized that his initial experience was not some strange anomaly. In his lab, he performed DNA sequencing, measured chemical composition, and did everything else more or less the way he does it today. It was the same way it had been done for over a century.
So Grøn enlisted his two co-founders, Daniel Storgaard and Magnus Madsen, and set out to build a full-stack platform with high-resolution cameras and lighting to capture visual data points and time-lapses. We recorded how different components vary within a particular component. Experiments respond to the conditions in which they are placed.
under the hood
Reshape has developed proprietary AI models trained on in-house data from its own labs, which are useful for more common types of experiments, such as experiments involving fungal and bacterial hosts, seeds and insects. It works from the beginning. But the company can also help customers train models for specific use cases, such as tracking how specific microorganisms behave under specific conditions.
“The Reshape data science team uses a custom-built MLops architecture to start from understanding the desired output and quantification, annotate the large required datasets, develop and benchmark models, and We handle model deployment end-to-end. It's a product for our customers,” Gron said.
For example, agricultural companies can use Reshape to test seed germination rates or the severity of certain diseases. Alternatively, food companies can perform ingredient characterization to test quality, freshness, and how ingredients ripen over time. This is usually anything that requires visual evaluation.
Assay detected proliferation Image credit: Reshape
Some of Reshape's customers are using platform technology to transition from chemical to biopesticides. Essentially, we're identifying which new compounds are most effective and documenting how they were made. And ultimately, speed is the biggest attraction for customers.
“They'll be doing four to 10 times more experiments than they were before, which means they'll be able to get products to market much faster,” Gron said.
Although Reshape allows results to be viewed in a cloud-based interface, the platform also supports data export in formats such as LIMS and CSV, allowing users to easily integrate with other biotech software such as Benchling or even just Excel. You can import data.
Results are displayed through a cloud-based interface Image credit: Reshape
In terms of accuracy, Grøn said the underlying model is compared to human performance in the same experiment, covering metrics such as false negatives. This helps avoid a scenario where an experiment might have been stopped prematurely because scientists thought it was ineffective.
“We are helping to reduce false negatives by about 80%,” Gron said. “We also help our customers reduce the time it takes to get results. Also, instead of relying on memorizing how you did an experiment years ago, we We track it completely. So every time we run an experiment on the platform, we track it. Reproducibility is very important.”
From a business model perspective, Reshape sells the entire platform, including the hardware, machine learning, and underlying software, as a subscription. Prices are charged based on a “value-based” pricing model, which varies by customer.
At this time, Reshape only ships machines in one size. In other words, if a customer does a large number of experiments, they need to obtain a large number of machines. Therefore, Reshape may require a larger machine to scale this up to large-scale industrial-grade experiments. Gron was a bit reluctant on the subject, but hinted that the company could “expand” into larger devices in the future.
Reshape Imaging Machine Image Credit: Reshape
growth
Reshape, a Winter 2021 graduate of Y Combinator (YC), has amassed a pretty impressive roster of clients, including Swiss agricultural technology giant Syngenta and the University of Oxford. Reshape, which has another $20 million in the bank following an $8.1 million seed round last year, said it plans to use the new capital injection to expand its U.S. operations, which account for about a third of its revenue. The second one is said to have already come out of the United States. Most of that comes from European customers' U.S. facilities.
“We've proven that our technology works, and now it's time to scale it and help as many labs as possible accelerate biological transitions,” Gron said.
Other companies are also bringing automation to scientific labs, such as London-based Automata, which raised $40 million last year to target a broader range of lab workflows. There are also companies offering services similar to what Reshape is trying to do, such as Singer Instruments' Phenobooth and Interscience's ScanStation.
But by offering a full-stack platform with end-to-end data management out of the box, Grøn believes this is what sets Reshape apart.
“This is an expensive problem that many companies have been trying to solve for a long time,” Gron said. “We provide incubation, image capture, and analysis in a closed-loop system. Our pre-trained models are ready to use right out of the box, so no time-consuming training is required.”
Reshape's Series A round was led by European VC firm Astanor Ventures, with participation from YC, R7, ACME, 21stBio, and Unity co-founder Nicholas Francis.