More robots means fewer lab coats. Image source: Ginkgo Bioworks.
Investors are going crazy for everything and anything associated with the Internet of Things, or machine-to-machine communication. After all, the field has the potential to add significant efficiency gains throughout the global economy in applications ranging from household appliances to agriculture to industrial manufacturing. While it is rarely discussed, the Internet of Things also has the potential to revolutionize biotechnology laboratories and research and development efforts -- perhaps creating substantially more value for biotech than in any other headline-grabbing applications.
The leaders in automated biotech lab services might not be household names today, but they're racing to build next-generation biofabrication facilities that will be difficult, if not impossible (or, at the very least, very expensive) to compete with. In other words, the early movers could be the only players in the market for quite some time, which could force future biotech companies to adopt business models strikingly different from those observed today.
What makes the Internet of Things incredibly important for biotech? What, exactly, does automated and interconnected biotech research look like? And what does it mean for companies such as Intrexon (PGEN 0.76%) and Amyris (AMRS)? Let's look at how automation and the Internet of Things are upending decades-old standards in the biotech industry.
Biotech's oldest problem, solved?
The biotech facilities of the future will rely increasingly on automation. Why? Consider what a typical day in a traditional biotech research lab looks like. Researchers sit at the bench, furiously hand write detailed parameters of their experiment (temperature, pH, oxygen content, and the like), tediously move small volumes in and out of test tubes hundreds of times, and track labels for dozens or hundreds (or more) of samples.
One big problem: humans make mistakes. They write down the wrong parameter (or forget altogether), add the wrong liquid to the wrong test tube, and mislabel samples. Even when an experiment is performed perfectly, two labs can produce very different results. Reproducibility is biotech's oldest problem.
Image source: Laurence Livermore/Flickr.
In 2011, healthcare leader Bayer took a random sample of drug development studies published in peer reviewed journals, while Amgen did the same in 2012. Each company performed the same experiments outlined in the studies in their state-of-the-art facilities to see if they could produce the same published result. Bayer found that no more than 25% of the results could be reproduced, while Amgen could only replicate results 11% of the time.
That's pathetic -- and costly. It is cited as a leading reason for high drug failure rates in healthcare applications of biotech and has decimated early industrial biotech companies such as Amyris that attempted to scale chemical production from 2-liter bioreactors to their 200,000-liter commercial scale counterparts. It highlights the finicky nature of biology (in our current limited understanding) and that humans simply aren't the best way to conduct biotech research.
Robots and software, on the other hand, always log the correct parameters, never add the wrong volume or liquid to a test tube, and don't mislabel samples. (Unless, of course, a human programs them incorrectly.) Only once biotech research is truly reproducible can we quickly commercialize new inventions and pinpoint errors in the R&D process. How is the Internet of Things coming to the rescue?
What does the future of biotech look like?
Robots. Lots and lots of robots.
Organism company Ginkgo Bioworks recently launched Bioworks1, an 18,000-square foot facility with 20 robots armed and ready to conduct biotech research. Everything in the facility is given a bar-coded label and logged in a virtual database. In addition to making it easier to follow each sample and step in an experiment, bar codes help robots to track inventory and automatically place orders for out-of-stock components to ensure research can continue without a hitch.
Image Source: Ginkgo Bioworks.
The company leverages its R&D expertise to build new organisms capable of manufacturing the cultured ingredients its customers desire, thereby monetizing the R&D process through contracts, and collects royalties if the ingredients are successfully commercialized. Customers save money by avoiding the need to conduct in-house research and gain a competitive advantage through a more effective or lower cost ingredient. Ginkgo Bioworks gets paid to compound its understanding of biology and widen the gap between internal and external capabilities, perhaps one day making it absurd for products companies to build their own organisms. After all, technology companies today don't produce their own silicon chips. We can thank Intel for that.
Similar versions of that business model are employed by other organism companies. Intrexon ended 2014 boasting 17 exclusive channel collaborations, or ECCs, that leverage the company's technology platform and automated R&D labs. While its future is heavily dependent on healthcare applications, CEO R.J. Kirk expects to generate $100 million in revenue from engineered cow embryos and other bovine reproductive tools and services in 2015.
Image source: Intrexon.
You can probably see where this is going. Automation is now a necessity for successful outcomes in biology-derived products, although there are no guarantees. Amyris learned some difficult lessons about reproducibility when it first attempted to scale manufacturing in 2012. While the company appears to be on the right path to deliver on its original potential, enabled in part by its talking robots (which I recently visited) that build 120,000 unique yeast strains each month, execution will make or break the investment opportunity.
Those early troubles weren't fun for shareholders, but eventually led to the departure of several key employees who would go on to start their own companies taking aim at biotech's reproducibility problem. Zymergen has quickly ascended the ranks as a leading organism company, while software start-up Riffyn aims to make biotech R&D truly reproducible through data collection, talking machines, and data analysis.
In addition to organism companies, several computer-aided manufacturing, or bioCAM, companies have built automated R&D platforms accessible through the cloud. Researchers residing anywhere with an Internet connection -- be it an academic lab, a major pharmaceutical company, or a local coffee shop -- can submit experiments virtually to Emerald Cloud Lab or Transcriptic, have armies of robots execute the experiment, and send an email when the results are ready for download. It's cheaper, the results are reproducible, and researchers can spend more time doing actual science.
Source: Emerald Cloud Lab.
While making biotech R&D more reproducible and predictable is great for corporations and academic institutions, automated platforms also put world-class infrastructure into the hands of the masses. We're nearing the point where anyone with a little biology knowledge, a laptop, and an Internet connection can create a custom organism or biology-derived product from a coffee shop. In fact, it's already possible today -- engineered houseplants included.
What does it mean for investors?
There are many implications for faster, better, cheaper, and more reproducible biotech research spanning industries ranging from healthcare to chemical manufacturing, food to consumer applications. The days when millions or billions of dollars were required to research, develop, and launch a biotech product to market will pass sooner than many think.
While many companies building and leveraging biotech's new, automated infrastructure are still accessing private capital, the bio-Internet of Things is already beginning to trickle down to publicly traded companies through cost savings. Sooner than later, individual investors will have no shortage of investment opportunities to choose from. Unfortunately, for now, more patience is required, but the future looks bright.
