The new era for AI-powered drug discovery and longevity biotechnology begins for Insilico

Alex Zhavoronkov, PhD
InsilicoMedicine
Published in
11 min readJun 11, 2018

It has been a long time since I published the Ageless Generation so my writing skills have become quite rusty. My group is publishing approximately two research papers per month now and many people have asked me to start a Medium blog to explain what we do at Insilico Medicine and Longenesis. Many of my friends are now using Medium and I should probably keep up with the trend. This is my first Medium post, so don’t judge it too harshly.

And this is a very good day to publish this post. Insilico Medicine just closed a strategic round of funding with four very strategic investors: Pavilion Capital, BOLD Capital Partners and Juvenescence. We also announced a strategic research collaboration with WuXi AppTec.

These are not just investors — each of these companies is a highly desired strategic partner with expertise in drug development and connections in the pharmaceutical industry.

These two events open a new chapter in my life and our company’s history. Since 2014, we have been working on a fully-integrated pharmaceutical R&D pipeline with many successes and failures along the way. And now, we have the resources to hire a business development team and perform even more laboratory experiments. The pharmaceutical industry is elitist, so I am looking for people with years of experience in the pharmaceutical industry capable of generating and closing Excientia-Sanofi-type deals at the C-level in the target identification and novel molecule generation space. We currently have 52 scientists, hired through hackathons and competitions in 8 countries, and now I am looking for a team to commercialize the fruits of our hard work. If you know someone who could win a “business development and licensing hackathon”, has a stellar resume with some pharma experience and would be able to co-pilot our rocket, I am open to suggestions. Please send them to our semi-automated assistant.

For chemists and biologists, please check MolHack.com or AgeNet.Net regularly and take part in many of our other online and offline competitions.

I am extremely thankful to my colleagues who have supported me during the good times and the bad. There were a few times in our company’s history when we went so low on gas that I had to sell everything I owned and put it into the company and they had to work extra hard to win the trust of our pharmaceutical, nutraceutical and academic partners. You guys are the best and I am very proud of you!

I am also specifically grateful to Dr. Ge Li, Mr. Edward Hu, Dr. Steve Yang, as well as their colleagues at WuXi AppTec. Ten years ago, I invested in WuXi AppTec when it was traded on the NYSE because I thought that they had a bright future. And today, it is truly a miracle research platform powering the entire healthcare industry. They are a dream partner when it comes to drug discovery, and I am deeply honored to be directly interfacing with their elite medicinal and synthetic chemistry team and their biology team. This investment is the result of a lengthy collaboration and experimental project related to our GAN-RL molecular generators. And my team is now committed and inspired to work on a universal system for generating novel chemistry for orphan biological targets with few or no training sets. This is the goal we set for ourselves for the near term.

Another deep bow goes to team Juvenescence, which led our round last July and gave us support and expertise that was impossible to find anywhere else. Juvenescence is focused on developing the next generation of drugs and other therapies for aging and age-related diseases. Learning the ropes of finance from Jim Mellon and drug development from the giants like Dr. Declan Doogan (developed Lipitor, Zoloft and Viagra), Dr. Gregory Bailey (funded Medivation and many other bioventures) and their A+ experienced drug hunters and drug developers and getting advice and support from experienced and amazingly organized executives like Anthony Chow is a rare privilege. Juvenescence is building a longevity empire and we are very happy to play a role in this process.

I highly recommend Juvenescence, the book, which details the longevity-centric investment philosophy and provides a good overview of the company and developments in this space.

I am deeply honored to receive an investment from BOLD Capital Partners co-founded by Peter Diamandis, the exponential entrepreneur who co-founded a large number of transformative businesses, educational institutions and high-tech high-impact non-profit organizations. In 2017, I had the pleasure of presenting at Exponential Medicine organized by Daniel Kraft and part of Singularity University, which Peter co-founded. Many of our team members read his books Abundance and BOLD. Our seed investor, who I also would like to thank here, Dmitry Kaminskiy is an early proponent of these books and at Deep Knowledge Ventures these books are a required reading. BOLD should be the required reading at every pharma company. I am deeply thankful to Neal Bhadkamkar and his wonderful team for believing in us and for his powerful support and guidance.

Another investor group which I would like to thank is Pavilion Capital. I am grateful for their strategic support. Pavilion It is a subsidiary of Temasek Holdings, a national wealth fund of the Government of Singapore. Thanks to Pavilion I will be going to Singapore more than once this year and will accelerate our expansion into Asia. Pavilion has experts in artificial intelligence and healthcare and is one of the most desired investors to have on the register bridging Asia and the West.

What is Insilico Medicine?

Very often I am asked what Insilico Medicine does. The answer is as simple as it is complex: we are an artificial intelligence company developing a comprehensive end-to-end drug discovery pipeline to extend healthy productive longevity for everyone on the planet.

Aging Research is at the Core of Everything We Do. And I personally think aging research must be:

  • Credible. We need to publish in peer-reviewed journals and work with the top research institutions, pharmaceutical companies and clinical centers exceeding normativeethical standards. To win trust among ourpeers and the general public we need to go the extra mile to comply with rigorous academic peer-review and the FDA approval process.
  • Goal-oriented. Age-related diseases and aging do not wait and many patients are suffering. We cannot do the research for the sake of research, it needs to result in products as quickly as possible. We need to develop biomarkers and interventions with results/outcomes that people can see, believe i, that regulators can approve and healthcare systems can ultimately deploy.
  • Sustainable. The personal computer, the internet, smart phones and social networks did not gain mass appeal by relying on government grants. A for-profit company must focus on creating shareholder value and achieving sustainable growth and revenue.
  • Presentable. Bill Gates, Steve Jobs, Elon Musk, Mark Zuckerberg, Jack Ma, Jeff Bezos, Jensen Huang, Sergey Brin and Larry Page did not build their businesses into the global conglomerates they are today by being silent nerdy actors. They were fit, young, confident and outstanding speakers and communicators. Longevity biotechnology needs these effective communicators and Jim Mellon and Eric Verdin are among the best examples to follow.

We try to follow these principles at Insilico Medicine and establish a solid model and reputation working within the traditional framework of the pharmaceutical drug discovery and development.

Longevity as a Service

My long-term goal is to develop a fully-automated Health as a Service (HaaS) / Longevity as a Service (LaaS) engine, which can plug into Alibaba, Tencent, Amazon, Facebook, Alphabet and Apple etc. and deliver personalized solutions that will help prevent a variety of diseases and maintain the users of these conglomerates in their optimal health state. I hope to cover this concept in more detail in my follow-on posts and describe just a few lego blocks we are developing on the path to this grand goal.

It is clear that the investment thesis of this round is not aging. It is the validation of our GAN-RL pipelines. To learn more about the GANs for drug discovery, I recommend these focused Medium posts by a wonderful friend, collaborator and advisor and a column on GANs covering some of our work. My Chinese readers may like the recent article in MIT Technology Review China.

In brief, we are using GANs to ‘imagine’ novel molecular structures with specified molecular properties (over 20 different properties) and reinforcement learning (RL) to generate the molecules to a specific objective. We started this journey in 2015 and submitted the first proof of concept paper to a credible journal in June 2016. Since then we spent two years doing theoretical and experimental work with varying success, and only recently we got the techniques to work at a level we were satisfied with. We usually publish the results that are 12+ months old to show the direction and update the community but internally the generative pipeline is quite advanced. It is also a reasonably small part of the big pharma.AI pipeline.

And unlike many other companies, we are trying to build the entire “salami” while publishing only small pieces of it. And we bring modern AI solutions to pretty much every part of the pharmaceutical R&D pipeline as well as for the real world insights and even marketing:

Our goal is to make such a great disease-target-molecule combo that it would go through clinical trials with flying colors. It will take time.

And while my favorite system (which I use almost every day) is Young.AI and the Pharmacognitive.com system available for the general public contains $1.6 trillion worth of grant data, what drives our company is a drug discovery pipeline. Most of the many Lego blocks there are seamlessly integrated and automated.

Insilico Medicine Drug Discovery Pipeline

And it pains me to say that the target identification pipelines, pathway reconstruction and causality inference engines are probably more developed than the generative pipeline. But the beauty of our approach is that we are integrating the two pipelines with many scoring mechanisms in both the biology and chemistry domains.

The predictors of clinical trials is a very new initiative stemming from a paper we published in 2016, subsequently abandoned, but later resurrected at the request of a VC firm which looked at us for six months and later started a company of their own (possibly more on this in later posts). Clinical trial predictors are very difficult to develop from available omics or structural data because a great many of them fail due to poor design, financial or political decisions, or were just strategically shelved or abandoned. And the molecules that passed clinical trials may not necessarily be great drugs by today’s standards.

One great thing, is that we recently combined a really cool set of new databases with new data types that no one previously thought of combining, and found some really unusual features that are likely to be affecting the probability of success even when the technology changes from small molecules tomAbs to gene therapies or something completely novel we have not yet thought of. We are not sure if we will be able to build an accurate predictor, but even what we have today could be quite useful for anyone interested in portfolio management of research megafunds.

So, what I can tell you about our pipeline is that it looks like this, and in about a month we will allow you to try our pathway analysis tools as a service for your own drug discovery:

One of the areas we pioneered was deep-learned predictors of human chronological age using pretty much every biologically-relevant data type. These techniques relate to our target ID pipeline, where we reconstruct the pathways from the most relevant features and establish causality. But some of these networks and features can be used as markers of biological aging.

And while many scientists think that they know the value of the data and know what data will be driving the progress in the future, believe me, they do not have a clue. Nobody does. We are just scratching the surface of the life data economics on the macro level (omics) and micro level (chemistry) and just like in finance we need to find a way to value the data.

That is why through our JV with BitFury called Longenesis we partnered with several groups working in this area and one amazing group focusing on the application of blockchain technology — Nebula Genomics. Please look it up and consider piloting our joint platforms and collaborating on understanding the value of human data in the context of clinical trials, but also in the context of health and longevity.

Conclusion

Pretty much everyone knows that I am very dedicated to extending human productive longevity in every way shape and form. I may not be a great scientist or a great business person, but since 2004 I have worked 18–20-hour days to understand the nature of aging and disease to identify therapeutic interventions that can generate more quality years of life and make humans more resilient to all kinds of stress.

I have taken this chance to tell you the brief story of Insilico Medicine, and describe our roller coaster ride to where we are today, and it has been a wild ride indeed. In the last two years, the company ran out of funding a couple of times and I genuinely believed that the company waswere finished. On one of these occasions, I had to sell pretty much everything I had and even borrow to add a few vital months to ensure that the company survived and retainmy top talent.

With the funding we have now, I will be able to grow the team, hire senior leadership and perform the experimental validation of our compounds to take us to the next level.

Please consider following us on PubMed:

And please consider attending some of our conferences especially our forums in Basel

It is a very happy day and if you are at the CogX conference, the Financial Times @ Cognition X or at the Founders Forum this week, please visit our panels and talks and let’s find a way to collaborate, to make this world a younger and healthier place to live.

--

--

Alex Zhavoronkov, PhD
InsilicoMedicine

Founder of Insilico Medicine, adj. prof at the Buck. I speak at 40+ events per year on AI for drug discovery and longevity research. PubMed: Zhavoronkov+A