BIOS Podcast #6: Computational Biology w/ Nan Li — Managing Director @ Obvious Ventures
BIOS: Nucleus of Life Science Innovation 🚀
- Community: World’s Largest #TechBio Community — Apply to Join 🎉
- Write: Share Thought Leadership — Become a Guest Contributor 🔬
- Newsletter: Subscribe for Exclusive Insights — Click Here 🔥
Nan Li is a Managing Director @ Obvious Ventures, where he invests in companies solving the world’s biggest problems in sustainable systems, healthy living, and people power. At Obvious, his investments have included LabGenius, Recursion Pharmaceuticals, Darwin AI, Octave Health, Zymergen, and Planet.
Prior to joining Obvious, Nan has had a mix of technology, investing, and entrepreneurship experiences including early stage tech investments at Innovation Endeavors, Product, Operations, and Finance at Gigwalk, VC and management consulting at Bain, and PM at Microsoft. Nan is also an adjunct lecturer at Stanford where he teaches a course on symbolic systems in VC and entrepreneurship.
This article is a summary of key takeaways from the BIOS Podcast episode w/ Nan Li — Listen here!
Computational Biology Investments
Nan credits his successful computational biology investments to having a prepared mind and being at the right place at the right time. Prior to Obvious, Nan invested at Innovation Endeavors, where the firm had the mandate to invest cutting edge technology before the term “deep tech” was coined. During that time, it was increasingly clear that the success of Illumina and the decreasing cost of sequencing would a treasure trove of data in biology.
As a result, bioinformatics was a topic of interest for the firm and Nan had the good fortune to meet the Zymergen team in 2013. The company was 6 people at the time, but painted a very clear picture of how data science and machine learning could accelerate the synthetic biology process. The architecture that Zymergen wanted to create was a platform where biological experimentation and machine learning fed into each other and formed a virtuous cycle. Having a front row seat as the company actually built this out made him even more interested and he realized that the primitives formed at Zymergen could be applied to a whole host of biological problems.
Understanding the Biology of Investments
“It’s very daunting! Important to stick to what you know, especially within biology since there are so many nooks and crannies. It is important to be a lifelong learner.”
Nan stresses to stick to what you know, which for him was CS, ML, and data science approaches, which he augmented with a vocabulary and understanding of very core biological concepts. Bringing something to a table that resonates with founders and having biological awareness of latest developments to contextualize conversation was key to allowing him to lean heavily on his strengths in computation while still providing value as a VC.
“Ultimately, looking at very complex companies, the core philosophical insight or mechanism is quite simple. There can be a crutch in making things as complicated as possible. Sharp clear minded founders can distill their company to different levels. It is a negative sign if you’re drowning in PhD level conversation without core insight. The core insight of the company should be simple and clear. Not many issues here. If there are, it is a red flag.”
Applying Digital Approaches to Biology
The convergence of software and digital approaches with biology and life sciences brings together two camps with very different core processes. To understand this convergence, Nan classifies these using the dichotomy of platform centric models vs asset centric models. The question here is whether a platform in biology can generate multiple assets and produce a pipeline of different products, or whether drug discovery is capturing lightning in a bottle. In pharmaceuticals and traditional life science, there is a sense of mysticism.
A term prevalent in the industry to describe a good scientist is a ‘drug hunter’. To Nan, this says a lot. It shows that drug discovery is not procedural. Rather, it is a black box, a result of luck and some soft skills. The new wave of tech companies entering biotech don’t take that approach. Nothing in CS rivals that level of mysticism. New companies are trying to build procedures and productized platforms and approaches to systematize or engineer biology. That can be a really challenging or heretical idea to the traditional life sciences model of innovation and drug discovery.
What about the skeptics about computational drug design?
“Any new technology should trigger some amount of skepticism from the industry. This is entirely warranted. The whole idea of computational drug design used to be hypothetical, but now there are clinical trials you can point to to prove that it can work. Recursion has 4 clinical stage assets that are in patients today. Those were discovered by their platform which is highly computational in approach. So you are seeing this translation from philosophical underpinnings to tractable results which have made the transition to the vocabulary and currency of the industry: advancement of assets, maturity of pipelines, and clinical results. The category of computational biology or AI for drug discovery is getting validated more and more every day.”
Where will we be 10 years from now?
Nan notes that any interesting cycle in tech takes a lot longer than VCs or entrepreneurs think; their worldviews are usually too fast paced. Something as large as the convergence of tech and pharma may take many decades. However, Nan thinks that we are entering a golden era of biological innovation and collaboration.
“In 10 years we will be able to speak of many more successes, and this toolkit will be applied towards a much broader set of indications. Just like AI has swept into enterprise, almost not worth mentioning that a company is using AI, the same thing will happen here where the tech will be inserted across the industry. Here we are still a bit defensive in pointing to case studies. In 10 years the industry will change to sharing best practices or methods that other companies can use or adopt. Less focused on point solutions that are leading the charge.”
Along the way, the underlying infrastructure that powers computational biology will also advance (robotic automation, methods for data management and machine learning) ultimately producing a richer and more robust set of assets that will emerge into the clinic and be commercialized.
Managing Fund Cycle Expectations in Deep Tech Investing
Building a company of consequence takes time, and this is especially true for any heavy duty industry like pharmaceuticals. For Nan and the Obvious Ventures team, the entire point of venture capital is to be patient capital. Venture got its start by funding challenging long term projects and Nan believes that those are the use cases where venture should exist.
“We want to be patient investors and watch a company grow over a decade. We set those expectations with LPs. They know that these companies take time, contextual awareness, and research. On the startup side, watching a company grow from seed all the way to large category leader will take over a decade. Not only in terms of FDA, but building a platform, and also finding the right assets to advance.”
However, the feedback loop is actually much tighter in practice. Nan notes that he was constantly learning about what the Zymergen platform was capable of. With the story arc of Zymergen growing through the last 7 years, now it is very clear to everyone that it has potential externally, but for the Obvious team, it was clear from the beginning.
“We were able to see day to day what the platform was capable of. So as long as we have this conviction and closed loop empirical evidence that these platforms are working, coupled with LP support and the disposition to be patient, we were totally aligned to watch our companies and not hurrying to exit. We wanted to make sure they take their time and stay in the private market to ensure that they take the time to mature.”
Paths Through the Clinic: Manifest Destiny vs Platform + Partner
For companies that are deciding on whether to take their assets all the way through the clinic or to partner with large biopharmas for distribution, Nan offers the following advice:
“If you believe that you are building a platform that can generate multiple assets and dont have one piece of IP that you have to protect, but rather a method or mechanism that can bear fruit over a long time, it makes sense to be long term oriented. It makes sense to partner out the first asset or give up economics on early products, as a tradeoff for industry buy-in, signal, or to make the company more capital efficient. Those tradeoffs make a lot of sense if you have multiple assets.
I’ve given the advice not to be short term greedy and focus on the economics of the first deal. I’ve told founders to think about 2nd order effects and play long term games. If you can build the platform on someone else’s dime, go ahead and do that even if you’re giving up some of the economics of the first deal. The partner approach is something platform companies should at least explore to stay capital efficient. It also helps to start industry relationships and over time to help contextualize the tech platform against a real case study from industry.”
East Coast vs West Coast Investment Mindsets
Tech forward bio companies on the West Coast think about platforms in a very different way than East Coast platform companies. East Coast giants like Moderna or Juno are platform companies, having multiple assets, a core biological mechanism that they believe in and are commercializing across multiple indications. But these are not tech enabled platforms.
When Nan meets new companies, he asks them: “Are you advancing commercializing a core biological insight or a tech / digital insight”
“If you believe that you found a mechanism of a disease or invented a new class of vaccines, these are biological insights that can generate platforms. However, the West coast companies are not coming out of labs with bio insight.
Their insight is about systems and application of data science in these problems. Data science explores biological questions in a way that interrogates biology without creating and reforming testable biological hypotheses. The way that computational biology works is to rapidly move through the synbio states where you are designing a set of experiments, building them out, running and testing the experiments in the wet lab (biological space) and taking the results of those tests back into digital space to learn and generate next step of experiments. That loop with ML in the middle is the core of computational biology. If you can run through that cycle fast and cheap enough, that can evolve the answer faster than thinking about and rationalizing the problem to find a solution.”
However, Nan cautions that this framework hinges on the quality of the test phase. Teams are only as good as the assay, the speed of experiments, the cost of each experiment, and the quality of signal capture from each experiment. Those three are atomic unit of the whole company and extrapolation of those characteristics makes up the quality of the platform, the capital needs and all of the other properties of a comp bio company. Nan likes to dive in and characterize that atomic unit because he sees it as the founding box for any company.
Day to Day Advising Roles
The model that is practiced at Obvious is very high touch investing and as a result the team invests in very few companies relative to other firms. Nan spends a lot of his time with his companies rather than finding more companies. On a day to day level, it means spending enough time learning so that he is informed as an investor about the space, about the companies he works with and what they are trying to achieve.
“I think bio investing is just a category of investing that is not well suited for tourists. Tech + bio is becoming more popular but ultimately these companies still have the uncertainty and nonlinearity of biology in the company. It is different from software where you know how long it will take to build something, and you just have to sell it. These companies are different, there’s an unknown time element. I try hard to stay up to date on research, current events, and what is going on in academia in the spaces I invest in so that we can have real conversations about what’s happening, and about how where they fit in.”
Nan notes that having a context allows the team at Obvious to have conviction through the ups and down and to not use intellectual shortcuts to measure progress like revenues, partnerships, or other vanity metrics.
“I ultimately want to know whether the platform is working or not, how well they are working and in what spaces. That’s the true measure of progress. If that holds true and continues to grow, there will be lots of commercialization options if the platform works. I spend a lot of my time working with founders to get to the ground truth about platform operations, about learning about their core technology, trying to make them feel like they are supported and that they don’t have undue pressure to sign first dollars of revenue before they are ready or to sign a partnership, when they actually can’t live up to the contract. As investors, the best piece of support is to make them believe that we know what they are trying to build, that we support them for the long run, and that our support is unwavering in small up and down blips as the company matures. To get that level of buy in and trust, it takes a lot of over communication both ways.
I have a casual cadence with companies for this reason. I don’t care about board meetings, I just want to know what is happening all the time. We can bypass the formality and just talk about learnings and adjustments and operating the company more on a week to week basis than big milestone investor updates. This approach makes founders comfortable with being long term oriented and not super focused on just delivering one great quarter.”
Learnings from Early Investments
Nan believes that we are at the beginning of an important transformation. However, the most important thing that has come up repeatedly is that even though these companies are pushing forward a radical new way of doing things, interfacing with pharma and speaking the same language of the industry is still very important.
“If you’re trying to develop novel materials or therapeutics, you need the help of the industry, you have to ultimately leverage a pharma company’s resources to bring your idea to life. The disruption model in life science is collaborative by nature. You will never see a Netflix blockbuster story here. For the companies I’ve invested in, the founding idea is to do something different and there’s definitely the entrepreneurial spirit, but they’ve balanced that with awareness of the industry that they are going into. This applies even for simple things like what assays resonate more when talking to partners. Being able to wear two different hats, where one it’s all about innovation and doing something different and thinking outside the box, and two you have to take all of that and package it so that it looks inside the box. Otherwise I think it’s very difficult to work with industry and get the partnerships and capital that you need.”
Ultimately being able to be a startup and to build something radical is important but it’s equally important to not do it in isolation and to not alienate the industry where you really need external support. Some of the biggest accomplishments are not only what the company is able to build internally from a tech standpoint, the external support to bring their products to market.
“It’s important to think about this early on and not just at the end when you want to tack on a sale.”
Where is this next revolution going?
At a high level, Nan is excited by opening up the search place in material and therapeutics from where they were historically. The rise of Zymergen and other biomaterials companies have unlocked prior bottlenecks in target driven drug discovery or undruggable targets have been opened up a bit by computational search and navigation through system and engineering biology.
“When you have a hammer of this magnitude, all you see are nails, in every therapeutic modality. It’s important to understand mechanistically how it works, what the assays will be, what the experiments will be, but I believe that there are many areas where you will see new companies develop that take this philosophy and mechanism to build verticalized companies that are comp bio focused.”
Some of the verticals that Nan has been paying special attention to where there aren’t any big companies but certainly a lot of activity and interest to pharma are CAR-T and T-cell activation, inflammation and immunotherapy, and manufacturing downstream post discovery.
Advice to Early Deep Tech Founders
It’s really important to stay true to your core capability and what you are trying to accomplish instead of form fitting to certain archetypes. I’ve seen companies force themselves to be a platform when really they only have one asset to commercialize. I’ve also seen some companies that are forcing on a computational or AI component when their core insight is a biological mechanism. There’s nothing wrong with that, there’s tons of ways to build great companies no matter what path you take so I think it’s better to overrepresent what you are and what makes sense for you and to find the right investor for that. Force fitting yourself to each investor that you see will come across as inauthentic and I think it’s very confusing not only for investors but also your co-founders. The best advice I can give is to know what game you’re playing, have the confidence to play it, and to not deviate from it just because of market trends.
The other thing I would mention is that in biology because progress is unevenly distributed and somewhat nonlinear, founders need to really be thoughtful about building a company and to think about multistage, multi round planning. For companies that raise a seed, they should think about what progress needs to be made to raise the next round and about how much of that progress is under your control and how much of it is still driven by biological discovery. Certain biologically based binary events need to happen and founders should give themselves a buffer and do scenario planning against that. I think to build a deep tech company in general, especially a biology company, requires a different planning than building a software or any product based company where you are in control of your own development schedule. Even companies I work with, we think a lot about having a buffer for unknown timelines and making sure that we are extra conservative about getting past a certain threshold for progress before going to market. It’s worth putting in extra effort upfront so you don’t get caught in a valley between two peaks without proper runway or cash management. It’s an issue that I see come up a lot and it’s not an interesting issue, you won’t see a lot of public press or thought leadership around it, but I see it as super important.
Advice For 10 Years Ago
I don’t know what time period this is most applicable and I would actually give this advice to myself today. At least in the current venture environment, I think that venture is becoming more specialized, which is a reaction to the industry becoming more mature and saturated. If I look at starting my career 10 years ago and how I represent myself and now, I’d certainly begin to tunnel in certain areas where I have strong conviction over the next decade or longer versus looking at everything and trying to be a generalist. I think that specialist networks and knowledge, contextual awareness, is critical to being a truly helpful investor to founders versus dishing out generic business building advice. I can see it even in my own career where I invested in the first company in a category and now in comp bio I’m investing in my 7th or 8th company. I truly think that you become better if you specialize. So for anyone who is thinking about getting into venture, I would spend some time thinking about how I can become a top investor in a more narrow segment versus a very active investor across a large number of segments.
📣 If you enjoyed this post please clap 👏 & comment 💬 to let us know
Join BIOS Community 🎉
Become a member, continue the conversation, connect with like-minded Life Science innovators, access exclusive resources, & invite-only events…
Apply to Join — Membership Application