on the assay side, the molecular biology side, in terms of the enzymes being used for sequencing. They are being engineered to be better. The reagents are being designed to be more pure. So the error rate is decreasing, and the quality of the chemistry and biochemistry is improving at the same time as the throughput of the instruments is increasing.
X: How far ahead do you think the technology is getting beyond the biology? Do people have any idea what to do with all these sequences? What kind of experiments does this open up for biology?
DW: This is kind of the great debate between hypothesis-driven research, where you have specific scientific questions you ask, and correlation science, in which you gather the data first and then see if you can figure out what’s going on. We’re at the stage where there’s enough capability in genotyping and sequencing, and we’re on the cusp of seeing these two approaches come together. A lot of hypotheses out there in biology will be addressed by being able to put these high-throughput techniques to work.
The original hope was that you’d do these genome-wide association studies, and you’d understand the molecular basis of things like Type 2 diabetes, and understand lupus, lung cancer, breast cancer and everything else. It turns out these are multigenic diseases. There are likely to be many different factors in play, spread widely across the genome, that lead to complex interactions and result in disease. A single mutation such as sickle-cell anemia, for example, with a point mutation, is not the common type of genetic disease. Usually it’s a matter of multiple sites, and multiple variants that play a role in disease. Through the sequencing efforts, we’ll discover many of the rare SNPs [single nucleotide polymorphisms]. In the next round of Genome Wide Association Studies (GWAS) we’re going to learn a lot more. The next round of papers isn’t going to say, ‘This particular SNP was responsible for 8 percent increase in risk for Type 2 diabetes.’ It’s going to say, ‘Here’s the map of Type 2 diabetes,’ which we’ll begin to understand by genotyping hundreds of thousands or millions of people. We’re going to begin to understand the real genetic underpinnings of disease.
That doesn’t eliminate the fact that’s there’s another layer of complexity involved, and that’s environmental factors. You have to look at the context of the environment in which your genetic makeup is immersed. People who live in New Orleans are probably going to be more prone to obesity, because they’re eating rich, fatty foods, than, say, people who live in California. They’re immersed in an environment where such factors will have a stronger influence.
X: So when people say we’re going to sequence all the genomes and it’s going to lead to personalized medicine, it sounds like you’re saying, ‘Hold on, this is going to take longer, it’s more complicated.” Is that right?
DW: Over the next couple of years, we’re going to know a lot more than we ever did before. The expectation back in 2000 when the completion of the Human Genome Project was announced, was that biology was solved and we would understand all these diseases. That expectation was a very simplistic view of things.
These discoveries will be a 50 to 100-year kind of thing. We’re going to continue to learn about human disease very broadly. But with all these tools, you asked if we’re getting ahead of the biology. We’re already ahead of the biology. The data being generated are way ahead of the biology. But as long as we don’t lose that data, as we get to 1,000 genomes, 10,000 genomes, 100,000 genomes, or 1 million sequenced, then we’ll be able to use sophisticated tools to mine the data. Eventually we’ll be able to get the answers to some of the key questions.
X: So computing will be the key?
DW: Yes, that’s the limitation. The bioinformatics side is not there yet.
X: So do you have a breakthrough idea for a new company?
DW: We started a new company called Quanterix. It’s based in Cambridge. I can walk to it. It’s focused on single molecule protein detection for early diagnosis. Basically you’re able to measure the levels of proteins at unprecedented levels of sensitivity. One example would be to find very early markers of cancer in blood. You might have a 1 millimeter-cubed tumor that doesn’t pose any threat, but by taking a blood sample you may be able to find proteins it is producing.
X: Maybe you would put people on chemo earlier then?
DW: Maybe, or maybe just monitoring and seeing if the immune system takes care of it and you don’t have to worry about it. But if it continues to grow, you put people on chemo earlier. Then you can see if you’ve done anything effective or not.
X: Could this be the next Illumina?
DW: I hope so.
