clock menu more-arrow no yes mobile

Filed under:

Google Ventures' Bill Maris on Moving Medicine Out of the Dark Ages

Google Ventures is on the hunt for companies using computer science to improve health.


Google Ventures made health startups a sizable part of its portfolio from the earliest days, even as other investors avoided the space in recent years.

Venture capital funding for the life sciences sector dropped by $5 billion from 2008 to 2012 and was basically flat last year, according to market reports. But the search giant’s venture arm, established in 2009, has steadily plugged money into companies throughout the space, including: 23andMe, Adimab, DNANexus, Doctor on Demand, Foundation Medicine, Flatiron Health, iPierian, One Medical Group, Predilytics, Rani Therapeutics, SynapDx and Transcriptic.

Some of the bets have started to pay off. Foundation Medicine raised $100 million in an initial public offering in 2013. Earlier this year, Bristol-Myers Squibb bought portfolio company iPierian in a deal that could be worth up to $725 million.

The focus on the space at least in part reflects the background of Google Ventures’ Managing Partner Bill Maris. He studied neuroscience at Middlebury College and neurobiology at Duke University. In his early career, he was the health care portfolio manager at Swedish investment firm Investor AB.

Maris also took a lead role in the creation of Calico late last year, a Google-backed company focused on delaying aging and the diseases that come with it. (Google has declined to discuss the company, which is run by Genentech Chairman Arthur Levinson.)

Google Ventures generally isn’t taking the old biotech route, betting on companies somewhere along the winding path of developing drugs that may — but probably won’t — someday earn Food and Drug Administration approval. Rather, the firm is focused on companies leveraging the increasingly powerful capacities of computer science, including big data, cloud processing and genomic sequencing, to improve diagnostics or treatments.

In the second part of my two-part interview, which has been edited for space and clarity, Maris discusses the promise of these tools for medicine as well as what’s still standing in the way.

Re/code: Looking through your health-care investments, there’s 23andMe, DNA Nexus, Foundation Medicine, Flatiron. To the degree there’s a common theme, it seems these are all big data plays, using a lot of information and smart algorithms to make advancements in medical research or hit upon more effective treatments. Is that part of your investment philosophy?

Maris: I used to be a health-care investor a long time ago in the public markets. One thing I learned that we tried to apply here is that investing in small molecules, trying to invest in the next treatment, there’s an element of gambling to that.

I’m glad that people started those companies and I’m glad that they have people who specialize in investing in them. But that’s not our specialty, because you have to build a portfolio to make a success overall.

What we try to put into our practice is “invest in what we know,” which is where health care meets technology. In some sense, almost all companies these days need to be big data companies.

Especially when you get around genomics or, like Flatiron, looking for insights across vast amounts of oncology data. These are by definition big data companies that couldn’t have existed 10 or 15 years ago.

Take Foundation Medicine. The tools didn’t exist to actually genotype quickly the way that we can today, and in 10 years it will be even more advanced. So by necessity the companies we’re investing in are in that space, because that’s the forefront.

Clinicians treating patients based on “if you present with these symptoms, I’m going to treat you based on the knowledge in my head?” Those days are either disappearing or will soon disappear, I hope. We can get much better outcomes from people if we understand the genetic basis of the exact cancer that they have, what interventions might be most effective against it, what’s worked in the past and what hasn’t. I think that’s where the future of health care is.

So yes, lots of these are big data companies, in that sense. But that’s a catchphrase, they’re more than that. They’re data-informed companies that are trying to build businesses that are commercially important and, in this case, relevant to patients. That means they’ll get better outcomes, you’ll live longer and be healthier.

Medicine needs to come out of the Dark Ages now.

There is a unique challenge when it comes to data and medicine. Either you have a lot of information that is stored away in paper filing cabinets in doctors’ offices, or you’ve got companies that did studies decades ago that might be of use but they’re either not digitized or they’re holding on to them as intellectual property. So while there’s this great potential, it’s actually really hard to get at it. Can you talk a bit about what needs to happen technologically?

Of course it’s difficult. If it were easy it would be done by now, there would be nothing remarkable about what Nat [Turner] and Zach [Weinberg] are doing at Flatiron. The fact that it’s difficult is what makes it something an entrepreneur needs to tackle — and this isn’t unique, right?

All the information in the world has been pretty dispersed, but Google’s mission has been to organize it and make it universally accessible. That’s kind of a crazy mission and they’re doing okay at it. It takes people with a vision to say, “We’re going to try to organize this and make it accessible to people.” When we do those things, good things will result from that.

Maybe it takes a generation, because doctors will start using the system. Or maybe it just takes one big push, where we’re just going to go into clinicians’ offices and help them get all the data organized and put into electronic formats. Once you’ve done it one time you can gain an infinite number of insights to help your patients, so there’s a good motivation to do that.

Organizing healthcare information is a daunting task, but it is not an impossible task. We’ve had people walk on the moon. This is a lot more doable.

I want to ask about 23andMe. We’ve seen a handful of companies in that space that have closed or haven’t gone anywhere, and 23andMe obviously hit a big wall with the FDA last year.

I don’t know what you’re talking about.

Yeah, I read it somewhere. But that was a big part of their business, can you talk about what their ongoing prospects are and what direction they could steer in?

Yeah, as I understand it, the heredity product is still available and we see big businesses being built there, like and others.

At the same time, their vision is bigger than that. They’re at an impasse with the FDA right now, but no one has thrown up their hands. Communication is ongoing, they’re trying to work that out, we’re dedicated to trying to resolve that roadblock. And we think it’s a product that is of value to people, so they can look at and understand their own genomic information.

I think the company’s prospects are great, I’ve known [co-founder] Anne [Wojcicki] for almost 20 years now, and she’s nothing if not focused, dedicated and motivated. She’s a believer in this. I think the company has been a little bit ahead of its time.

It’s inevitable that everyone will eventually be genetically sequenced because it’s going to be really important to their health care, to understanding their future and what they’re at risk for. If you believe that, then you believe that there’s probably a big business to be built here because someone has to deliver that information.

So we have a lot of faith in the team.

Taking that case — and given that health care and medical research is moving in this digital direction — do you think there are some regulatory shifts that need to take place?

I think the laws need to catch up with science and reality, and the law is never good at that. It’s always slow.

I mean, look at the patent office. I just saw a patent that Smucker’s has for a peanut butter and jelly sandwich. It’s sort of crazy.

Look at Uber and its regulatory challenges, taxi and limousine commissions trying to stop Uber. When you sit with my job — which is a really fun job to do, kind of a judge at a science fair — it’s really important to look at the technology and how it might benefit people, and not worry about the bureaucracies that might try to impede that.

At the end of the day, what always happens is, the right products for society and the people get out there.

You shouldn’t ignore the laws. But if you worry as an investor about, “Oh, you shouldn’t invest in any personal genomics companies because there’s a lot of regulations that need to be updated.” Well, you won’t do anything innovative.

So yes, absolutely, the regulations need to catch up with reality. I think as the outcomes of the science with Foundation Medicine, 23andMe, etc., start to become important to people and to patients, people will demand that change. And that’s how it happens.

You studied neuroscience and neurobiology. What are some exciting developments you’re seeing in your own area?

I also think we’re just coming out of these Dark Ages in neuroscience. The forefront of neuroscience is (he points to parts of his head), “Well, this is the learning area, this is memory, this is where the right arm is controlled.” That’s not really how the brain works, it’s this cloud-based understanding.

I forget which neuroscientist said this, but you essentially have a Jennifer Aniston neuron. There are certain pathways in your brain that remember who that is. The more you fill up your brain with those things, the more neurons get used up.

So we’re getting closer to a point, and there are some folks at MIT working on this and other places as well, to really understanding the wiring of the brain. What makes it a whole, what causes consciousness. It’s not just that these cloudy regions all talk to each other.

You can’t do anything without a map. Until you can diagnose something you can never cure it, you can’t understand it. It’s hard to get from here to there without a map. So the first thing to do is to build a model.

When you can map an entire human brain, then you can really understand how it all works.

We don’t even know if everything gets recorded in your brain and your brain is just really good at controlling noise, where it’s just filtering out a bunch of things that you don’t need to think about because you’d just be overloaded. So there are these fundamental questions of neuroscience we just now have the tools to understand.

It’s so far behind, it’s so underfunded, in a way. We as a people and a country spend a lot of money on a lot of things. But we all walk around with this thing in our head and we have no understanding of how it actually works.

Machine-brain interfaces are a way to understand that. There’s a guy at Duke named Miguel Nicolelis, who I worked with and who comes out here every once in a while. He does work where he implants electrodes into brains and he’s now got monkeys who can move cursors on a screen [with virtual arms] and they get a reward of orange juice. Then he thought, “Well, why is the monkey just limited to one [virtual arm]? Maybe I could teach them to move three at once, or four.”

What we are learning from that is, well, we have two legs and two arms, but your brain is actually capable of operating four or six of them if you had them. There’s so much potential.

Here’s what the monkey saw in that experiment:

This article originally appeared on

Sign up for the newsletter Today, Explained

Understand the world with a daily explainer plus the most compelling stories of the day.