9 Computational Drug Discovery Startups Using AI
Recently we talked before how big data is the new frontier with just 0.05% of all data available today having been analyzed. There are really two reasons why this number is so low:
- 98% of big data has only been created in the last several years.
- A meaningful amount of this big data is the incessant drivel that you see people vomiting all over every form of social media (data which is turning out to be surprisingly useful in predicting whether or not they’ll pay back a loan).
This means that all kinds of gold prospectors are lining up with their freshly crafted artificial intelligence (AI) algorithms looking to extract all the value they can from this wild west of data before someone else does. Perhaps nowhere is there more excitement at the moment than the applications to be had in the healthcare industry. Here’s a look at just some of the startups that are applying artificial intelligence and big data to healthcare (courtesy of the bright minds over at CB Insights):
The application that we’ve circled above is “drug discovery” using AI or what’s also known as “computational drug discovery“. The reason that this is now a thing is not just because of all the big data that’s available now, but also because of how cheap cloud computing has become, not to mention the emergence of deep learning algorithms. Earlier last year we threw together a list of 4 startups playing in this space, and later came across another called BenevolentAI that made our list of the 5 biggest AI startups. Now we want to take an updated look at the players in this list and focus on what makes them different.
In order to help us do that, Andrew Radin, co-founder of TwoXAR, kindly offered his assistance. If you recall from our last article, Andrew M. Radin was one of the two “Andrew Radins” who founded TwoXAR which explains the startup’s name. Yes, the two actually met while quarreling over www.andrewradin.com which is maybe the coolest “how you got started” explanation ever. Andrew M. helped explain the drug discovery process in the simplest of ways as follows:
- Find a new protein in body to hit with molecule
- Find molecule(s) that binds to protein in body
- Once you find a hit, then turn into something that can be introduced to a living being
When it comes to computational drug discovery, a startup can focus on one or many of these steps. Let’s take a closer look at 9 startups involved in computational drug discovery.
The most valuable of the bunch is British unicorn BenevolentAI which has taken in $100 million in funding at a valuation ($1.78 billion) that makes them the largest AI startup in Europe and one of the top 5 biggest in the world. With a life science paper being published every 30 seconds and an FDA approval process that is all but broken, BenevolentAI plans to use AI to speed up the drug discovery process such that they plan to “sell their own drugs directly in the next 4 years“. An article published yesterday by Business Insider says that BenevolentAI “signed an $800 million deal in 2014 to hand over two Alzheimer drug targets to an unnamed US company for development” and that since they started in 2014, they now have “24 drug candidates“. So yeah, seems like pretty promising stuff. Nerds like us will enjoy the unique way in which they chose the name for their first clinical study.
Founded in 2007, San Bruno California startup Numerate has taken in over $40 million in funding from investors that included Eli Lily to develop a computational platform that can predict how a drug will behave in the body. Taken from their website “at the start of a typical program, we virtually assay 25 million compounds from a bespoke, focused virtual library of 1 trillion (or more) compounds, against a handful of accurate activity, selectivity and ADME models at a cost of one-one hundredth of a penny per compound, in about one week”. The process is fast enough to search through spaces of 50 trillion compounds in one week (on 10,000 CPUs).
Founded in 2013, Salt Lake City startup Recursion Pharmaceuticals has taken in $15.35 million in funding to support their ambitious goal of finding 100 disease treatments in 10 years. Instead of using the inefficient strategy of studying an explicit molecular target related to a specific disease, they use high-throughput biology, advanced imaging, and artificial intelligence to make many discoveries in parallel. Incredibly, they actually use computer vision to look at cells and identify over 1,000 features that can be used to determine if a sick cell is getting healthier when exposed to 1000s of drug compounds. They’ve partnered with big pharma firm Sanofi, and have a candidate going into clinical trials this year according to an article in MIT Technology Review.
Founded in 2014, Baltimore Maryland startup Insilico Medicine has taken in $10 million in funding so far to tackle aging and age related diseases. Aging and telomeres have been a hot topic lately here on Nanalyze, and Insilico wants to tackle not only aging but also cancer. Don’t we all. Similar to NuMedii, the startup looks at drugs that are already safe to use and see if they can be re-purposed for other uses. Eventually, they’d like to conduct full body digital simulations which kind of sounds like what the most funded AI startup in China is doing, iCarbonX.
Insilico works as a “contract research organization” and they’re using Nvidia GPUs and machine learning algorithms to make this all happen. The AI division of Insilico is called “Pharmaceutical Artificial Intelligence” and they not only have a very nice futuristic website but they also “take pride in keeping our identities away from professional and social networks” and they warn people to “use extreme caution when hiring any of the professionals claiming to work for Deep Pharma or Insilico Medicine“. The “war for talent” is getting a bit heated, no?
Founded in 2012, San Francisco startup Atomwise has taken in $6.35 million in total funding so far from the likes of Khosla Ventures and Draper Fisher Jurvetson. That money has been used to develop AtomNet, “the first structure-based deep convolutional neural network design to predict the bioactivity of small molecules for drug discovery applications“. There’s a pretty heavy white paper you can tuck into for details, but with 27 drug discovery projects launched already with leading research institutions, it sounds like their platform has definitely been validated for usefulness. If you tried to run their algorithms on a typical laptop they would take 10,000 years, but they’re working with IBM cloud and IBM Watson to make it happen at a fraction of the time. They’re also working on confidential projects with Merck and Autodesk as well.
Founded in 2008, Silicon Valley startup NuMedii has taken in $5.5 million in funding to commercialize “big data” licensed exclusively from Stanford University which consists of hundreds of millions of human, biological, pharmacological, and clinical data points that have been normalized and annotated. One of the biggest struggles facing AI companies is having to cleanse and transform big data sets so they are usable, so NuMedii is ahead of the game there. The startup mainly looks at “repurposing existing drugs for new indications“.
The youth of today should set aside their aspirations of being a dime-a-dozen “famous vlogger” (look it up) for a moment and check out what this ambitious young lady and man set about doing with their spare time. These two kids haven’t even completed their PhDs yet and they’re already starting a company with $4 million in seed funding from some big-name investors and have managed to get George Church himself as a Scientific Advisor. Verge wants to look at which FDA-approved drugs out there that have expired patents and can be used to treat neurodegenerative diseases. We’d like to see more Medium articles highlighting success stories like these two and less talking about all those first world problems we face every day.
Update 07/18/2018: Verge Genomics recently received additional funding of $32 million to be used in starting their programs on Parkinson’s disease and several others that their CEO, Alice Zhang, declined to name.
Founded in 2014, Silicon Valley startup twoXAR has taken in $3.4 million in funding so far with Andreessen Horowitz as the lead investor. The Company’s technology platform was inspired by work performed while one of the co-founders was studying Biomedical Informatics at Stanford University. twoXAR-developed AI-based algorithms are trained on billions of real world biomedical data points including gene expression measurements, protein interaction networks, and clinical records.
According to an article in Wired last month, twoXAR worked with Stanford to screen 25,000 potential drug candidates for liver cancer. It narrowed the list down to 10, the most promising of which is now headed toward human trials. The only existing FDA-approved treatment for that same cancer took five years to develop yet this recent candidate took twoXAR and Stanford just four months to find. Earlier this year twoXAR, signed an agreement with Santen, a world leader in the development of innovative ophthalmology treatments, to look for glaucoma drug candidates.
Founded in 2006, Berg Health is backed by Silicon Valley real estate billionaire Carl Berg and has taken in an undisclosed amount of funding so far. According to an article in The Economist, “the firm’s AI system starts by analyzing tissue samples genomics, clinical data relevant to a disease. It then tried to model from this information the network of protein interactions that underlie the disease“.
When thinking about the differences between these 9 computational drug discovery companies, it’s best to refer to the simple 3-steps of drug discovery we highlighted earlier. Companies like BenevolentAI, Verge, NuMedii , and TwoXAR are all taking a holistic approach and trying to re-imagine the entire drug discovery process to work as efficiently as possible. On the other hand, companies like Atomwise and Numerate start working at the second step. Once there is a known protein, they then figure out what molecule can bind to it and what will be safe. In some cases they create entirely new molecules. Then there are companies like Insilico that want to offer “computational drug discovery as a service” (CDDaaS??). Then of course there is Recursion which still blows our minds in that they actually look at cells using computer vision to see if they get well when treated with different compounds. What an amazing time to be alive.
Here at Nanalyze, we invest the majority of our savings in 30 dividend-paying stocks that increase our income every year and outperform during both bull and bear markets. Find out which ones in the Quantigence report freely available to Nanalyze subscribers.