C&EN’s 2017 10 Start-Ups to Watch

Anyone lucky enough to receive a new Lego set gets instructions on how to build the model on the cover of the box. But there is no step-by-step manual to guide entrepreneurs creating new businesses.

This year’s selection of 10 start-ups to watch were chosen by C&EN’s writers and editors because they have both innovative chemistry and the vision to use discoveries to solve important problems, targeting unmet needs in agriculture, chemical identification, drug discovery, materials design, and renewable products.

Not all the chemistry they use is brand-new. For example, lyotropic liquid crystals have been known for years, but the start-up Light Polymers has found a niche for the materials in improving digital displays. In agriculture, Boragen is exploiting boron chemistry first used in pharmaceuticals to craft highly targeted crop protection chemicals.

Other companies to keep an eye on are harnessing advances in artificial intelligence, analytics, and visualization. These developers of futuristic tools will make it possible to watch proteins change shape, identify molecules in the field, and develop new drugs and materials through experiments designed by machine-learning algorithms.

All start-ups face difficult odds. Statistics show that only 1.1% of start-ups grow to employ at least 50 people in their first 10 years. According to the National Venture Capital Association, venture capital investors are now backing fewer firms at later stages of growth.

The trick is to put the right bricks together at the right time. We think these 10 start-ups have the blueprints for success.

And if you know an exceptional new company based on chemistry, nominate it for next year’s feature at cenm.ag/startupnom.

Arrakis Therapeutics

Targeting RNA with small-molecule drugs

By Ryan Cross

Gilman is now CEO of that start-up, Waltham, Mass.-based Arrakis Therapeutics, which was founded in September 2015 by Jennifer Petter, former vice president of chemistry at Celgene and now chief scientific officer for Arrakis.

“It’s a bit like getting the band back together again,” says Gilman, who previously worked at Biogen along with Petter and Arrakis Chief Business Officer Daniel Koerwer in the late ’90s through mid-2000s.

Gilman has already helmed two successful start-ups: Stromedix, acquired by Biogen in 2012, and autoimmune-focused Padlock Therapeutics (featured in C&EN’s 10 Start-Ups to Watch in 2015), acquired by Bristol-Myers Squibb in 2016. This February, he helped rake in a $38 million series A funding round for Arrakis.

Many companies have tried targeting RNA in the past decade using techniques based on RNA interference, or RNAi, in which a strand of RNA binds to and helps eliminate a complementary strand in the body. It’s been difficult to get the technique to work well in humans, however, except for one recent success in treating liver disease.

Arrakis is attempting to block RNA with the same chemistries used in protein-binding drugs. Gilman thinks that’s possible because contrary to oversimplified illustrations, RNA is more intricate than a simple strand; it twists and folds around itself into structures that present promising binding spots for drugs. Gilman doesn’t expect all RNA molecules to selectively bind a drug, but he says that Arrakis wants to “find small molecules that stabilize one of these transient, dynamic structures of RNA and essentially lock it down.” The small molecule would thus prevent an mRNA from forming a protein or inhibit a noncoding RNA that controls gene expression.

“People will always ask, ‘Why haven’t we found selective RNA-binding compounds in the past?’ ” Petter says. “Well, you actually have to look.”

Koerwer says Arrakis hopes to make a name for itself by doing that searching. Arrakis has one disclosed program: targeting mRNA for huntingtin, which is the protein associated with Huntington’s disease. After that, Arrakis isn’t committing to any particular diseases. Instead, the company will “screen as many libraries as we can against hundreds and hundreds of targets, and then choose from that data set,” Koerwer says.

If Arrakis can find selective small mol­ecules that bind RNA and demonstrate a biological effect, Petter anticipates the firm will run drug development and clinical trials just like any other pharma company. “Our ambition is for the drugs to be painfully boring” compounds that can be packaged in pills and easily reach their targets in the body, Petter says.

Still, Arrakis must first demonstrate that it can be done. “This is going to be hard,” Gilman says. “But hard doesn’t scare me; that was actually part of the attraction.”

UPDATE
This story was updated on March 16, 2022, to correct Jennifer Petter's name.

Boragen

Protecting plants with shape-shifting boron molecules

By Melody M. Bomgardner

Liu was introduced to the idea of bioactive boron molecules—and his future Boragen cofounders—at the start of his postdoctoral research at Pennsylvania State University. He met scientists who successfully deployed boron in treatments for skin infections and other diseases, proving that boron can be used to make safe, biologically active molecules.

It turns out that agriculture has plenty of biological targets to pursue with the help of boron. That opportunity gave Boragen a spot at the new Research Triangle Park, N.C.-based AgTech Accelerator. Earlier this year, the company emerged from AgTech with $10 million in funding from accelerator investors including the Bill & Melinda Gates Foundation, Bayer, and Syngenta. Boragen’s compounds can control a range of fungal diseases in plants with a mode of action entirely different from that of traditional fungicides.

What’s more, boron compounds have low toxicity to non-target species and work synergistically with traditional fungicides, the firm says. By using both modes of action, farmers can prevent the fungal organism from rallying its means of chemical resistance.

Boragen’s lead compound is a benzoxaborole, which works by throwing a wrench in a cell’s protein-making machinery. It inhibits aminoacyl transfer RNA synthetases, enzymes responsible for attaching specific amino acids to tRNA. The resulting malformed protein doesn’t work the way it should, killing the pest.

To fine-tune the activity of boron pesticides, scientists take advantage of boron’s unusual shape-shifting ability, Liu says. Boron compounds can easily be made into a trigonal planar shape or a three-dimensional tetrahedron to better fit a target site. The two shapes allow researchers to select a negative or neutral charge.

Boragen now occupies greenhouse and laboratory space in Research Triangle Park, a hotbed of commercial agriculture research, and is testing its lead fungicide in field trials.

Liu and his growing team of boron chemists and plant scientists are still exploring the many biological pathways that boron compounds can interrupt. In addition to beating fungal disease, boron chemistry could be used to combat nematodes and livestock parasites.

“We are so excited and passionate about the boron atom,” Liu says. “It gives you flexibility that you can’t get with carbon.”

Citrine Informatics

Taking artificial intelligence to the materials research lab

By Rick Mullin

Launched in 2013, the company developed algorithms that allow computers to zero in on the chemical and physical laws underpinning materials research, Mulholland says. Its software was also designed to access a huge store of untapped data—information about failed experiments that is traditionally trashed. “These companies are driven by confirmation bias,” says Mulholland. “They celebrate a successful materials win, but when a material doesn’t work, they just throw that away as a failure.”

Materials science was ripe for AI, says Mulholland, looking back on his experience in commercial R&D. “We were very good at generating data about our processes and our materials, but we weren’t good at using it for anything. We didn’t do much with it other than look for pretty straightforward trends.” As a result, R&D was slower and more challenging than Mulholland felt it should have been.

He and his partners also sensed urgency, viewing the sector as on the cusp of changes that would require a leap forward in the lab. “Materials are about to take center stage in the next generation of products. The companies that will succeed will be the ones that can innovate most rapidly,” Mulholland says.

While developing ideas, the company began working with a research group at the University of California, Santa Barbara, which had put together a database on thermoelectric materials. The firm worked with other academic research groups and began marketing to industry in 2014.

Citrine isn’t naming customers, but in September, it reported that scientists from UCSB and a joint venture of General Motors and Boeing used its software to solve a long-standing challenge hindering the use of aluminum alloys in three-dimensional printing.

Citrine, which has eight Ph.D. materials chemists on its staff of 26, has completed two rounds of investment, most recently a $7.6 million series A led by Innovation Endeavors, Prelude Ventures, and Data Collective. The World Economic Forum selected the company as one of the 30 most promising technology pioneers for 2017, and Citrine won the World Materials Forum Start Up Challenge Award this year.

Mulholland says Citrine, which opened an office in Pittsburgh over the summer, has no plans to move beyond materials science, despite inquiries from other sectors, notably pharmaceuticals. “We respectfully decline,” he says, “because we are focused on advancing the next generation of advanced product development software in materials rather than for general chemistry.”

Ecovia Renewables

A new approach to designing biotransformations

By Alex Scott

Once a family of bugs has been engineered, the fermentation process is simple and can be fed with inexpensive raw materials, Minty says. He predicts that the commercial cost of running some of the firm’s fermentation-based processes could be almost two orders of magnitude cheaper than existing routes.

Ecovia is currently focusing on producing high-value industrial chemicals, including molecules for the personal care sector. It is working with a range of raw materials, including glycerin, an abundant by-product of biodiesel manufacturing.

The firm has a full-time staff of four. All employees—as well as an angel investor—own equity in the company. The goal is to add one or two more researchers by year-end to bolster Ecovia’s capability in downstream processing and polymer chemistry.

The firm’s approach has already attracted funding from U.S. government agencies. In April, for example, Ecovia disclosed it had been awarded a $750,000 grant by the National Science Foundation for a two-year program to develop a suite of superabsorbent polyglutamic acid-based biogels for hygiene applications, including diapers.

The biodegradable biogels will compete against synthetic polymers such as sodium polyacrylate. Success could lead to the production of fully biodegradable diapers and potentially help reduce millions of metric tons of waste every year, Minty says.

In June, the firm took a major step forward when it was chosen as one of three winners of a full partnership at AkzoNobel’s Imagine Chemistry technology competition for its polyglutamic acid-based biogels. The firm had already started developing a route to these biogels through the NSF-funded project. The AkzoNobel prize adds to this work by providing a joint development agreement to take a product to market introduction.

“Compared to our past awards, the recognition from AkzoNobel sends a strong, validating message about our business and the markets we are addressing,” Minty says. “We envision scaling up our process and launching our first products in two to three years.”

Getting a first product to market using a novel technology can be a major challenge. Ecovia is hoping its partnership with AkzoNobel will pave the way to success for its one-pot biotransformations.

Exscientia

Artificial intelligence cuts costs and speeds up drug discovery and design

By Ryan Cross

That program formed the basis of the now-five-year-old Exscientia, whose Latin-derived name means “from knowledge.” The company hasn’t bothered raising money from external investors, preferring to make deals directly with drug companies, including Sanofi and Glaxo­Smith­Kline. Hopkins says the program outlined in the Nature paper “only represents a very small part of our entire platform now.”

The design-make-test-repeat strategy employed by Exscientia’s algorithms should sound familiar to medicinal chemists.

First, Exscientia feeds its algorithms a hearty helping of data obtained from many sources, including the partner pharmaceutical company, scientific literature, and in-house surface plasmon resonance screening, which identifies fragments of drugs that have potential for binding a tough target (ACS Med. Chem. Lett. 2013, DOI: 10.1021/ml400312j).

The algorithm crunches the data to suggest promising compounds. Next, a chemist selects 20 realistic-to-make ones, which Exscientia’s pharma partner synthesizes and runs through a variety of screens. The results are returned to Exscientia, which embarks on another 20 compounds based on those results.

The whole cycle takes about two weeks, explains Chief Operating Officer Mark Swindells. “The speed is quite alien to some pharmaceutical companies that are happy to take two months,” he adds.

Werner Lanthaler, CEO of the German contract research firm Evotec, was intrigued and began a collaboration with Exscientia in April 2016. The two firms set out to develop a bispecific small molecule—one drug that binds two targets—to inhibit CD73 and A2AR, proteins implicated in immunosuppression of cancer.

“We picked something difficult to test-drive what Exscientia was claiming to bring to the table,” Lanthaler says. Happy with the early results, Evotec announced 18 months later that it would take an $18 million stake in Exscientia. “It is very rare that someone is overdelivering and underpromising,” Lanthaler says. “And that is the situation we have here.”

To establish itself as a designer of bispecific compounds, Exscientia also formed a pact with Sunovion to create bispecifics for psychiatric disorders. It inked a deal with Sanofi, worth up to $273 million, for discovering bispecifics that manage two aspects of metabolic disorders, such as glucose levels and weight loss. Exscientia also has a deal worth up to $43 million to deliver preclinical candidates for up to 10 targets selected by GSK.

Hopkins has a catchphrase to emphasize the computer-chemist collaboration and alleviate fear of AI-induced job loss: “This is as much about IA, or intelligence amplification, as it is about AI.”

Hopkins says the structures suggested by the algorithm are reminiscent of ones a human chemist would design. “That’s been a really important cultural breakthrough in getting human chemists and machines to work together.”

Light Polymers

Lyotropic liquid crystals for lighting and displays

By Alexander H. Tullo

The materials are also ideal for suspending particles like the phosphors used in LEDs. The suspension is used to make a thin film that aids in the down conversion of light, in which blue light is converted into white.

In conventional LEDs, the phosphors are suspended in silicone. Marc McCon­naughey, Light Polymers’ CEO, says lyotropic liquid crystals can suspend the phosphors more compactly and precisely than silicone can. The result is 25% greater conversion of light.

“You can reduce these optical and thermal losses, but more importantly, we can tune the film to emit to certain spectra,” he says.

Such tuning can make for high-quality lights like those used in museums. Or light can be tuned to the red frequencies that are becoming popular for skin treatments.

The company has plans to release a series of lighting products under the Crystallin brand. One of its first products, the GrowBlade for indoor horticulture, debuted at an agricultural technology event in New York City in September.

Another target for Light Polymers is polarizers for displays. Currently, the polarizer films in LCDs are made from polyvinyl alcohol that is dyed and stretched in a solvent-based process. Light Polymers wants to replace it with water-based chemistry. “Basically, we think we can change this $10 billion industry,” McConnaughey says.

Lyotropic liquid crystals can also be used in circular polarizers for OLEDs to prevent reflection of ambient room light. They can be coated directly onto the display without being made into a film.

Light Polymers’ technology originated in Russia. In the 1970s, the Soviet Union wanted to develop a homegrown display industry. In the late 1990s, many scientists who had worked on liquid crystals formed new companies in this field of research.

Morozov and Valery Kuzmin, now Light Polymers’ vice president of synthesis, were part of the original team working on this chemistry in Russia. After first working out of a research center’s labs in Moscow, they moved to California and then joined Light Polymers, which was founded in 2013.

The firm completed a $24.3 million series A financing round in March. In July, it raised an additional $5 million in an investment round led by Tsingda International Venture Capital and joined by the semiconductor equipment maker Tokyo Electron and the Japanese chemical firm JSR.

In October, Light Polymers agreed to form two joint ventures with the Tsinghua Innovation Center in Dongguan, China, which is supported by the government of Songshan Lake, Dongguan, and Tsinghua University. One venture is for the production of OLED polarizers and the other is for lighting.

P2 Science

Making ozonolysis safe for manufacturing

By Michael McCoy

One of P2’s founders is Paul Anastas, a synthetic organic chemist who directs the Center for Green Chemistry & Green Engineering at Yale and is often called the father of green chemistry. The firm’s other founder, and its chief science officer, is Patrick Foley, who was in the first class of Ph.D. graduates to come out of the green chemistry center.

The two scientists formed P2—named for Paul and Patrick—in 2011 to commercialize ideas coming out of the Yale program. Later that year the pair won initial seed funding from Yale-backed Elm Street Ventures. They also brought on Neil Burns, a veteran of specialty chemical companies including Pilot Chemical and Oxiteno, as CEO.

Central to P2’s strategy, Burns says, is its continuous-flow technology, patented with the European engineering firm Desmet Ballestra.

“Ozonolysis has been around for decades,” Burns notes. Like most specialty and fine chemical reactions, traditional ozonolysis is a batch process. It involves bubbling an O₃-air mixture into a large reactor over eight hours. The reaction is highly exothermic and can create unstable by-products. In P2’s continuous process, the volume of reactants is one one-thousandth of that in a batch process, and the reaction time is measured in minutes, not hours, Burns says.

P2 is initially targeting aldehyde and ketone fragrance chemicals made by reacting ozone with renewable fatty acids and terpenes. Development agreements with flavor and fragrance companies including Bedoukian Research and Symrise are yielding both new molecules and renewable versions of existing molecules now derived from petrochemicals, Burns says.

The firm has been producing pilot-scale quantities of these compounds at its Woodbridge, Conn., facility. In February, it completed a $9.6 million series B round of financing that was co-led by Xeraya Capital and the venture capital arm of the chemical giant BASF. Burns says P2 will use the money to build its first commercial plant, where its flavor and fragrance molecules will be turned out in larger quantities, and new products, such as surfactants, will be pursued.

At the new facility, P2 will also scale up a vegetable-oil-based route to azelaic acid. This versatile diacid is now made mainly by Emery Oleochemicals, which calls itself the world’s largest industrial producer and consumer of ozone, at a plant in Cincinnati.

Azelaic acid is used in acne and skin-whitening products as well as industrial goods such as polyesters and polyamides. Burns and his colleagues see potential to expand the market further with P2’s safe, continuous ozonolysis process.

Relay Therapeutics

Probing protein motion to tackle tough drug targets

By Lisa M. Jarvis

So what does it take to create a molecular movie? Relay exploits a multitude of recent advances. For example, when Murcko was mulling protein motion a decade ago, chemists commonly relied on X-ray crystallography to create three-dimensional images of a protein. But those pictures were developed at cryogenic temperatures; today, it’s possible to do crystallography at room temperature, better replicating how a protein exists in the real world.

Other advances include new screening technologies that detect the conformational changes a protein undergoes when a ligand docks to it; DNA-encoded libraries that provide a more diverse chemical palette than traditional small-molecule libraries; and computational power from partner D. E. Shaw Research that allows molecular motion simulations thousands of times as fast as previously possible.

Putting these tools together, Relay develops protein motion hypotheses that it uses to pursue targets when there is a clear line from a gene to a disease. Among the more notorious examples of impenetrable targets is the oncogene KRas, which is known to be mutated in a large swath of cancers.

“Sometimes the target is extremely well validated—like KRas. If only we could make a great drug against such a well-validated target, we’re pretty darn sure it’s going to provide a benefit for patients,” Murcko says. “Those are the kinds of targets we’re focused on.”

In the last year, Relay scientists have been able to create molecules that bind to proteins that have been “like Teflon” for the industry, Chief Executive Officer Sanjiv Patel says. The biotech firm’s two most advanced drug discovery campaigns are already in lead optimization, and the researchers have created chemical matter against one target that Patel says has never been drugged before.

The key isn’t accessing these technologies but figuring out how to use them efficiently, Patel says. “A big part of what this team has done is learn how best to deploy each technology—and the collection of technologies—to address the critical questions of each drug program.”

As it moves toward its first clinical studies, anticipated to begin in 2019, Relay is expanding its team, which should hit 50 people by the end of the year. The biotech firm is in the market for a new space that will give it room to expand into a fully integrated pharmaceutical company. “I’ve come to build a company that is sustainable and will be here for the long term,” Patel says.

Solid Power

Developing solid electrolytes for safer, high-energy batteries

By Alex Scott

Combining the electrolyte with a lithium metal anode can bring the energy density of Solid Power’s battery cells to 400–500 Wh/kg. This is more than twice the density of conventional liquid-electrolyte cells that use standard graphite anodes, says CEO and cofounder Doug Campbell.

Its two other cofounders, Conrad Stoldt and Sehee Lee, started the company in 2012 with the aim of developing cathode materials based on their research at CU Boulder. But in 2016, the firm switched to sulfide-based solid electrolytes after securing an exclusive license to commercialize technology developed by Oak Ridge National Laboratory.

Campbell is convinced Solid Power is developing the right technology at the right time. “The automotive industry is really starting to embrace the concept of solid-state batteries,” he says.

Historically, a problem with sulfide-based electrolytes has been the material’s brittleness, which can cause structural problems in a battery, according to Frank P. McGrogan, a Ph.D. candidate at Massachusetts Institute of Technology who published research on the topic earlier this year (Adv. Energy Mater. 2017, DOI: 10.1002/aenm.201602011). “You have to design around that knowledge,” McGrogan notes.

Solid Power says low-temperature processing of its electrolyte allows it to use polymer binders that mitigate brittleness. Meanwhile, the firm’s proprietary electrode and electrolyte separator chemistry and coating process ensure efficient transfer of lithium ions between electrodes, Campbell says.

A couple of months ago, Solid Power relocated to Louisville, Colo. There, the firm is starting to install what it calls prepilot production facilities where it can make hundreds of automotive-scale battery cells per month.

The firm recently initiated a series A financing round led by the battery manufacturer A123 Systems to raise $15 million to fund research and outfit the firm’s new battery assembly center.

Solid Power already has a multi-million-dollar joint development agreement with one European car company and is looking to add similar deals. Its goal is to get solid-state batteries into commercial electric cars in the next five to 10 years.

Solid electrolyte batteries can be very strong, and someday they could actually form part of a car’s structure. In the meantime, Solid Power is working to roll out solid-state cells for niche military and aerospace applications.

Two Pore Guys

Handheld nanopore sensor rivals accuracy of lab equipment

By Marc S. Reisch

Heller describes the battery-operated detector as a fast, “single-molecule-sensing platform with the accuracy, precision, and sensitivity of reference lab equipment” such as high-performance liquid chromatography/mass spectrometry instruments and fluorescence detection systems.

“We can cut out a lot of equipment, time, and cost with our device,” Heller says.

The basic technology for using nanopores as measuring tools has been around for 20 years, Heller acknowledges. Oxford Nanopore Technologies and Illumina, for instance, both use biological nanopores for DNA sequencing. Some of the technology to do that originated at UCSC, where Dunbar worked with nanopore pioneers David Deamer and Mark Akeson.

But Two Pore Guys’ technology, in part developed by a third founder—UCSC researcher Trevor Morin, who joined the firm in 2013—isn’t intended to sequence DNA, at least for now. Instead it is designed to screen for electronic signals from molecules such as disease-causing agents or environmental contaminants.

At the heart of Two Pore Guys’ detection system is a test strip etched with a microchannel leading to a silicon nitride chip with a 25- to 125-nm hole. The slide-plus-chip combo costs about 50 cents when made in low volumes; it could be much cheaper if made in high volumes in an electronics factory, Heller says.

The analysis gets under way when a user inserts the test strip into the handheld detector. A sample, placed on the reagent-containing strip, passes through the channels to the chip, where an electric current from the reader pulls molecules across the nanopore.

Apps loaded into the detector and specifically designed for each assay allow the reader to recognize and count the electronic signals from targets of interest as they pass through the nanopore. The detector transmits the raw data for interpretation to a smartphone or tablet. Results can be displayed on the detector’s screen.

The company—which has 70 employees, up from 50 earlier this year—is now ramping up device and test strip manufacturing, Heller says. “If you want to rapidly detect a molecule,” he says, “you’ll want our device.”