Looking at cancer as a complex system
Dr Anna Barker - Arizona State University, Tempe, USA
I’m going to be talking about cancer as a complex system, so what do I mean by that? Essentially we have spent decades, actually, understanding cancer at a very molecular level, around the genome, the transcriptome, the epigenome, you can name your ‘ome’ actually and we know how to do it. What we don’t think about yet is that cancer actually occurs in three dimensional human beings so we, we being complex systems thinkers, actually try to take that information that we learn at the molecular level and integrate it over scales so that we can begin to model a disease like cancer to fit what cancer actually is which is all of that information translated over time and in space. So it’s not an easy concept for a lot of biologists but more and more biologists are trained in mathematics and physics so they are beginning to understand that you have to go to the next step to actually understand cancer.
Why is this new CS model of looking at cancer so important?
It’s very important because to this point in time we’ve not yet understood how to really control cancer. So even though we have lots of treatments for cancer, once the cancer has metastasised we don’t do much better than we did fifty years ago, people still die from this disease. So we really don’t understand fully, if at all actually, why cancer cells metastasise, why nearly every treatment we give to a cancer patient causes resistance. It will only be understood ultimately in the context of a complex system.
Is this quite an intense and expensive process?
It’s the era of big data, as you probably heard in the meeting this morning, and cancer is information. The genome is digital information, it’s no different than a computer programme. So what we’re looking at now is this explosion of information because we’re sequencing all these genes, we’re looking at all aspects of the proteome, the genome, the epigenome, creating enormous amounts of data. We literally will probably sequence over a million genes this year, genomes actually. There are countries that are actually planning on sequencing their entire populations, if you can imagine that; even China has got a plan on the table to sequence a big portion of their population. So we’re creating a data tsunami of just unbelievable size but what we don’t know how to do yet is to analyse that data and put it together in a way that you can being to understand how is it going to inform the models that we need for complex systems. So the data is actually going to drive us much more rapidly to this area I’m talking about, complex systems, because yes you have a lot of numbers to crunch but you have to have a model to crunch them around. So that’s where complex systems models come in.
If this enables us to identify more subtypes of cancer does it raise more questions than it answers and ultimately cause more work?
That’s an interesting question because the way we’re actually sub-setting cancer patients now is we’re doing it around their genomes. We’re saying that, say lung cancer, for example, there are many different subtypes of lung cancer, breast cancer, almost every cancer. That’s aggregated mostly around pathways, in other words we work now by understanding that a gene is changed within a particular pathway and then that pathway, that gene, changes a whole series of other genes down that pathway. We don’t know a lot yet about how that actually works because essentially this is occurring, as I said, in three dimensional space over time so everything is in real time. So what we’re looking at is actually not in real time, we’re actually looking at everything that we’re looking at as though it’s stagnant, if you would, nothing’s moving. I think what you’re seeing, essentially, is this first generation of how we’re going to subtype cancer patients. I think it will be a lot more sophisticated than that in five years and I think we’ll move towards these different phenotypes, probably based on understanding the complexity of the systems actually where they’re expressed. So the genomes are going to be expressed differently depending on the individual and ultimately maybe groups of individuals who will share these various pathways. But it’s the expression of the pathways in terms of the phenotypes that will really make a difference. So if we get the same cancer, exactly the same cancer, if you go in for your genes to be sequenced and I go for my genes to be sequenced they will be quite different. So we’re like snowflakes essentially. We share some very interesting aspects of our biology but you have your own biology. So we’re going to have to understand that and you have to understand that in the context of the individual and you can’t do that just from the genome.
What are the concluding points of your talk?
What I would say is that we’re reaching a point now where we’re beginning to get enough information to actually synthesise and integrate it in ways and model it so that we can begin to understand and predict, for example, how to treat a cancer patient, when a cancer patient is going to become resistant. To do the kinds of things right now that are really stopping us from curing cancer and even from preventing cancer. So I think we’re now just beginning to be on that upward slope of gathering sufficient amounts of information about cancer that we can start to integrate it, model it and develop a whole new generation of tools that will probably lead us to actually cure some of this disease.