Skip to main content
Premium Trial:

Request an Annual Quote

Q&A: New York Genome Center's Harold Swerdlow on Impedance Matching in Sequencing


HaroldSwerdlow2_0.jpgName: Harold Swerdlow
Title: Vice president of technology innovation, New York Genome Center, since June
Experience and Education:
Head of research and development, Wellcome Trust Sanger Institute, 2008-2014
Chief technology officer, the Dolomite Centre, 2006-2007
Senior director of research, Solexa, 2000-2006
Unit coordinator and director of microarray core facility, Center for Genomics Research, Karolinska Institute, 1998-2000
Research assistant professor, University of Utah, 1998-1991
PhD in bioengineering and human genetics, University of Utah, 1991
BA in physics and mathematics, University of California, Santa Cruz, 1979

After six years as head of research and development
at the Wellcome Trust Sanger Institute in the UK, Harold Swerdlow recently joined the New York Genome Center as vice president of technology innovation. In his new role, he is in charge of developing new technologies for genomics research and, ultimately, clinical applications.

In Sequence recently met with Swerdlow at the NYGC to talk about what attracted him to the center and what types of technologies he and his team plan to develop. Below is an edited version of the conversation.

Why did you decide to come to the New York Genome Center?

It's the proximity of the Genome Center to all the hospitals and universities in New York. It's an incredibly powerful resource for pulling together a lot of different scientific, medical, and hopefully engineering talent into one place. The fact that the genome center has the institutional founding members, who are working closely with us, gives us a very big resource to pull from. We provide things for them, but they also provide samples, expertise, new disease systems, new questions that we can answer. Even one of those institutes, like Rockefeller or Columbia, has many very bright scientists working there. But having 12 different founding institutions, all contributing ideas and expertise into one place, is incredibly powerful.

The other thing that attracted me here was the fact that it's new, and I thought I could influence the structure, what science got done, what questions got asked, just establish something brand new and fun from the ground up. This place is one to three years old, depending on where you start the clock. We had an official opening about a year ago. We're making the rules up as we go and everything is changing constantly. We're hiring people like crazy. I have always liked startup ventures — when I joined Solexa, I was the third person. This place feels like a startup culture. It's also quite well funded; that's always attractive.

What is your role as vice president of technology innovation?

Technology innovation is the research and development component of the center. The primary thing that we need to do is some really cool, new, innovative science, to invent new methods and new protocols, as well as test new things from companies and other places that come in before they go into production.

There is a service component as well. We will do things that are useful for the production team, for the institutional founding members. Let's say they want to be at the forefront of RNA sequencing, but they don't have the time to test which of the 24 different RNA sequencing protocols are the best. We can do some of that testing within my group.

How large is the technology innovation group?

It's just starting out. We're going to grow to about five or six people within this year, maybe a bit more, that's the short-term goal. I'm about to offer some positions; we've been interviewing a lot since I got here. In a standard genome center, it would probably be mostly molecular biologists, but I'm also hiring some engineers. It will probably be about half and half, engineers and molecular biologists or biochemists. I have one guy who has a degree in chemistry, so we can modify surfaces and change chemistries if we need to.

The engineers that we are about to make offers to have training in DNA. This has changed since I last hired engineers. In the old days, you had to teach them all the biology. But now, some of them even have sequencing experience, they can do PCR, they understand what DNA is.

What types of technologies are you planning to develop here, both for research and for possible clinical applications?

I'm not sure my group will be involved a lot in clinical development; we'll see. Technology for clinical use needs to be locked down at an early stage, so it tends to be the older protocols that you are pushing now into the clinic, rather than the brand-new stuff that nobody trusts. That's true in the clinic, in forensics, and in some other markets where you don't want to be really on the cutting edge because you want something that people believe is reliable and reproducible.

We will definitely be looking into single-cell work. Exactly where that's going to go from an applications point of view I can't really say at the moment. Another theme that we're going to pursue is microfluidics, which is going back to my history. We are going to be looking at how you can do library preps and use smaller volumes, and what advantage that has. And then we're going to try to marry these two things, single cells and microfluidics, and see what we can do with that.

Generally, we'll be working on smaller volume library preps, using smaller sample amounts and samples without a lot of material in them. Whereas a conventional production lab might want a microgram of DNA, a single cell has six picograms of DNA, so the methods to handle that are very different. We're going to start pushing into smaller volumes to do that more efficiently.

The other aspect to that is, if you get good at handling small amounts of DNA, you can look at things like tumor DNA in blood, circulating tumor DNA. You can also look at circulating tumor cells, maybe at the single-cell level, and at non-invasive prenatal testing.

Would your work extend to clinical samples?

Sure, but in the short term it would be research samples rather than clinical samples. Everything we do at the Genome Center has a clinical bent. We're not about agriculture, we're not about, at the moment, looking at pathogens. Everything is really pretty tightly tied with human disease and wellness. We don’t have a mouse program, and I don't work with animals at the moment. That's not to say we'll never look at some mouse models but we definitely have a strong focus on human genomics and human genetics.

A lot of companies are developing microfluidics for genomics, as well as single-cell techniques. Where do you see the role of an academic center like the NYGC versus companies in developing such technologies?

I don't say that I'm competing with companies. I'm putting together a team that's maybe different than some companies. Companies always have to push the bottom line, so they are going for the low-hanging fruit. One of the beauties of doing this at the Genome Center is that we can work on things that may be a little bit less commercially focused, maybe not the lowest hanging fruit.

One example would be glioblastoma. Not particularly my lab, but the Genome Center is looking at glioblastoma at the moment, which is not the most popular cancer. It's not exactly an orphan disease but the big drug companies and the big players are going after lung cancer, breast cancer, prostate cancer, pancreatic cancer — the big ones. Glioblastoma affects a lot fewer people but is a devastating disease nonetheless. We have the opportunity, because we are not commercially funded, to go after the rarer things. That aspect of it would apply as well in the technology innovations space — that we can go after things that maybe don't have a blockbuster commercial future but are nonetheless scientifically interesting and very important to work on. They would not attract a small company that's just trying to pay the bills and thus are going to look for early, quick revenues. Some of the single-cell work we could do might be in slightly more esoteric systems that other people might not be working on.

How do you collaborate with companies?

We do early access, certainly. We haven't announced any formal collaborations with biotech companies, only with IBM, but we're definitely open to collaborate with biotechs. There are some obvious players in microfluidics we are talking to. I can't say more about it at the moment because we have not made any agreements. We don't want to do everything ourselves, so collaborations are definitely important for a small group.

Where do you see sequencing technology going in the next five years or so?

Part of why I'm interested in microfluidics is because I think the next big challenge is not so much in the sequencing platform itself but in the integration on both ends. The bottleneck in sequencing at the moment is feeding the sequencers and getting the data off and interpreted.

Illumina has done a really good job of making sequencing cheaper, but I think there is still a lot that could be done on the analysis side but also in feeding the sequencers. These things are very hungry. With the HiSeq X Ten, we can sequence 50 genomes a day, and those are whole genomes. If and when Illumina opens it up to other samples, like exomes and panels of genes, the ability to prepare enough samples to feed those kinds of instruments becomes the bottleneck. That's why I am generally interested in library prep and sample prep as a way to create a more continuous platform.

If your sequencing happens in three days but your data analysis happens in 10 days, that's not very good. If your library prep is three weeks and your sequencing is three days, that doesn't make sense. In physics, we call it impedance matching, where you are basically trying to make sure that all the parts of your system can talk to each other and are compatible to each other. You want to get the times and the effort and the cost for all the different steps to match up, so you don't have the library prep costing 90 percent and the sequencing only costing 10 percent.

For pathogens, for panels of genes, the cost of doing sample prep far exceeds the sequencing. Again, there is an impedance mismatch. We can sequence bacteria for $10 but the library prep costs $40. You can't really cram more and more sequence onto the same chip, so you have to reduce the cost of doing the library prep, and the only way to do that is to use fewer reagents, smaller volume.

Generally, I want to reduce the cost, increase the speed, and reduce the effort of doing library prep. We are working on, in a general way, improving these three things, and microfluidics is one of the tools we will probably be using, but it's not the only tool.

So what will sequencing technology look like five years from now?

"Integrated systems" is my answer to that question — "Soup to nuts," "sample to answer," whatever you want to call it, but you stick your blood in on one side and the answer comes out at the other side. And I mean a real answer, not three million variants, but something that a doctor can work with, which is one sheet of paper that basically says what drug should he or she should be prescribing and what they should be aware of for that particular patient.

In order to really make inroads into the hospitals and the point-of-care setting, we really need integrated systems. Whether that's microfluidic or not, it seems to me one machine that can do that is kind of where sequencing is going. I think that's going to be more important than whether sequencing drops from $1,000 to $500 to $250 per genome. Because if the bioinformatics costs $2,000 and the storage costs $2,000, what's the point of lowering the sequencing cost any further?