Skip to main content
Premium Trial:

Request an Annual Quote

Sifting Through Systems Biology


Dean Bottino, Novartis Institutes for BioMedical Research

Christopher Portier, chief, Laboratory of Computational Biology and Risk Analysis, NIEHS

Arnold Levine, director, Center for Systems Biology, Institute for Advanced Study, Princeton University

Andrea Califano, director of bioinformatics, Columbia University Genome Center


In late February, four systems biology experts braved a harsh winter (and 10 inches of snow) to gather at the Genome Technology office in New York City and hash out what they see as the meaning, future, and ways to measure success in systems biology. What follows is a transcript of their discussion, edited for space.


GT: Let’s start with how you would each define systems biology.


Levine: For me that’s become a complicated question. My answer is a cop out, but it’s true: I think it’s too early to say what systems biology really is and what it should be and what it will be. It has its broad outlines, like integrated biology — trying to integrate multiple signals from a large, noisy background in biological systems, be those signals from databases that collect information from the cell or from readouts of various kinds from the body or even from populations of people. It’s a field that will evolve into something that will be very different in five years.


Portier: I disagree. There is a clear definition in my mind of what I want systems biology to be and where I want it to go.

I’m very interested in knowing what happens when a chemical enters the body; I want to know from the molecular level clear up to the population level what happens. I want it well founded in data so that I can build a reasonable picture of why I see a particular toxicity. Taking that to a human situation and making predictions and projections of population risk is the name of the game and systems biology is one of the key components to take us there. With my definition I would argue that we’ve been doing systems biology for 20 years. All the work that has looked at population-based models and stochastic models of response, physiological and other types of differential equation type models … all combined with biochemistry and molecular biology type models now to give you a full picture of what’s happening in a system. It’s not just genomics, it’s not just proteomics; it’s the whole picture.


Bottino: Would you add that it’s not just biomedical or theoretical biology?

Portier: Absolutely. It’s got to be three things at a minimum: good biology, good mathematics and theoretical biomathematics, good biostatistics. Without all three components, you’re going to fail.


Bottino: One of the things I’ve noticed is the distinction between those fields, and I think people outside of mathematics are surprised to find out that most mathematicians either know statistics or know, say, deterministic modeling [or] differential equations — but rarely ever think to combine the two or become experts in the two. That’s been one of my challenges since I’ve gotten into industry is exploring that interface between deterministic modeling and the uncertainties that are so inherent in all the systems that we look at.


Levine: We all work on different things and we all like to think of ourselves as systems biologists. It’s a very diverse field and that’s exactly where it should be, because where it will be useful is unclear.

For a field to make an impact on biology, it really has to mature and get to the point where everyone in biology pays attention. Genetics is a great example.


Califano: Many ways that I’ve heard systems biology defined are reminiscent of a field we were already doing 30 years ago — electrophysiology — where we were creating models and simulations for biological systems. The fundamental difference today is the strong roots at the molecular biology level and particularly genomics.

The quality I find most attractive of systems biology is that it actually puts computational biologists and wet lab biologists to work on the same type of problems. The way we approached systems biology at Columbia is essentially by putting together a fairly large center for computational biology and the genome center.


Levine: You could almost mention a dozen schools now, or maybe two dozen, which are hiring physicists, statistical mathematicians, or computer scientists to be in biology departments or physics departments that do biology. In the extreme, which is just a wonderful example, Harvard Medical School created a department of systems biology. If Harvard Medical School created a department of systems biology, how far behind could the 122 other medical schools be? That’s a way to measure where the field is.


Portier: I’m a mathematician and a statistician, but I’ve had a wet lab for 10 years. To me that’s just an aspect of good science. And if you want to talk about extremes, I now run a $200 million research program as a biostatistician and a mathematician, so I have a lot of ways to change the way in which toxicology is done.


Califano: I think it’s too early to come up with a unique definition. If you talk to different people, they will interpret it in very, very different ways. Syd Brenner has this notion of the widgets in the cells, and Lee Hood has a completely different notion of how to deal with disease by looking at the mechanistic function rooted in the genome.


Bottino: Hopefully it’s driving a cultural change — that mathematical modeling isn’t something you do once you’ve figured out all the biology, but rather it’s a tool to help you organize your thoughts and to organize your next experiment and to guide exploration.

What I hope in terms of medicine that comes out of systems biology is that ultimately maybe there won’t be names for diseases anymore — but rather, this is this person’s biological systems, two or three of his parameters are out of whack and we can show why he’s not feeling well because of this; what combination of compounds or other interventions can we do to this person to restore proper function on a systems level?


Portier: I think one of the key issues you’re touching on is that biology is now too complicated to go the original route. Biology was a lot of rote memorization and learning the language — now the language is too complex and multidimensional. People recognize that, and the tool for organizing that information is in the mathematics/engineering/physics realm.


Levine: If that’s true, the people in this room are going to [make] obsolete a generation or two of people trained in classical biology.

It’s exactly equivalent to when I was young, of molecular biologists coming in and replacing biologists. There’s a potential here, but I still believe systems biology really has to prove itself before it’s going to be a replacement.


GT: In order to prove itself, the systems biology field is going to need some real success stories. Where do you see those happening right now?


Portier: There are big successes if you go to the broader definition of systems biology and avoid the narrow, genomic definition. There’s not a pharmaceutical outfit out there, nor is there a regulatory agency in the United States, that would move forward on a chemical or a pharmaceutical without having a pharmacokinetic model in hand.

We don’t do the study without having a model — that’s a classic approach to systems biology.


Califano: One way to measure success is in the rate of excellent publications that you see in journals. Right now if you look at papers in which computational science is involved that get published in the top journals, most of the really interesting papers that you see today are rooted in systems biology. This field is growing at an extraordinarily important pace.


GT: What are the major challenges facing systems biology?


Portier: Not a lot of people are able to think beyond four or five dimensions. You’re looking at manifold dimensions, so many things simultaneously — putting that into a framework that allows a human mind to understand and experience it is one of the biggest challenges that exist.


Levine: I actually think we have a very simple challenge: to become really good at what we do and really make an impact on biology. When three or four papers come out that everyone has to read and it really blows open the field and everyone starts doing it — that’s the challenge. Make a change the way molecular biology made a change.


Califano: I see two challenges. It is extraordinarily difficult to find biologists who do not in fact feel threatened by the computational scientists. Until there [is] a more level field where there will be a true appreciation of the different components of the life sciences, there’s going to be [a problem].

The other is the same exact challenge that was perceived when traditional biology was replaced by molecular biology. There was exactly the same sense of loss by one community.


Levine: It’s still not possible to publish a theoretical paper without data from a lab and get it accepted by a really good journal. That has to change.


Portier: Databases for models don’t exist. We don’t have a language for how we’re going to store them. We’re going to be wanting to link models from multiple datasets to understand entire systems.


Bottino: There are efforts in that direction, like SBML and CellML, but they’re not linked tightly. I don’t believe there’s a language that can explicitly describe what the perturbation was to generate a certain data set — that needs to be made more technical and rigorous as well.

I’ve seen this happen with buzzwords coming up and evolving and having a certain life cycle; one [challenge] would be to manage expectations. The other is, can we as a community be a little bit more careful about how we apply the term? Maybe we all have our own exclusive ideas of what systems biology is, but right now it’s getting sprinkled in like a magical pixie dust to improve the chance of publication or funding.

The Scan

Genetic Risk Factors for Hypertension Can Help Identify Those at Risk for Cardiovascular Disease

Genetically predicted high blood pressure risk is also associated with increased cardiovascular disease risk, a new JAMA Cardiology study says.

Circulating Tumor DNA Linked to Post-Treatment Relapse in Breast Cancer

Post-treatment detection of circulating tumor DNA may identify breast cancer patients who are more likely to relapse, a new JCO Precision Oncology study finds.

Genetics Influence Level of Depression Tied to Trauma Exposure, Study Finds

Researchers examine the interplay of trauma, genetics, and major depressive disorder in JAMA Psychiatry.

UCLA Team Reports Cost-Effective Liquid Biopsy Approach for Cancer Detection

The researchers report in Nature Communications that their liquid biopsy approach has high specificity in detecting all- and early-stage cancers.