Skip to main content
Premium Trial:

Request an Annual Quote

BioArray Q&A: Tufts' David Walt on the Evolution of Optical Sensor Arrays


By Justin Petrone

walt.jpgName: David Walt

Titles: Professor of Chemistry, Tufts University; professor, Howard Hugher Medical Institute; Founding scientist, director, and scientific advisory board chair, Illumina; Founding scientist, director, and SAB chair, Quanterix

Professional Background: 2004-present, professor, department of biomedical engineering, Tufts University; 1981-present, adjunct professor, department of biochemistry and pharmacology, Tufts University; 1992-1995, professor, department of chemistry, Tufts University; 1989-1996, chairman, department of chemistry, Tufts University; 1987-1988, visiting scientist, department of biology, Massachusetts Institute of Technology

Education: 1981 — postdoc, Massachusetts Institute of Technology; 1979 — PhD, chemical biology, State University of New York at Stony Brook; 1974 — BS, chemistry, University of Michigan at Ann Arbor

For nearly a quarter of a century, David Walt's laboratory at Tufts University has been applying micro- and nanotechnology to interrogate single molecules, analyze genetic variation, and investigate the contents and behavior of single cells.

Using fiber optic bundles, the Walt lab has specialized in developing optical sensor arrays that have served as the initial technologies for two companies — Illumina, which was founded in 1998 to commercialize bead arrays created in the Walt lab; and Quanterix, a Cambridge, Mass-based startup established two years ago to prepare the lab's single-molecule analysis technology for eventual clinical use.

A cofounder of both firms, Walt also chairs Illumina and Quanterix's scientific advisory boards. To learn more about the diverse projects taking place at the Walt lab, BioArray News spoke recently with Walt. Below is an edited transcript of that interview.

How does a guy from Detroit wind up designing optical sensor arrays?

When I first left Detroit, I went to Stony Brook. I received a PhD in what at the time was called chemical biology, but it's not what we call chemical biology today. It was basically a program that included both pharmacology and organic chemistry. Then I worked as a postdoc with George Whitesides while he was at MIT. I was working in the field of immobilized enzymes, which involved attaching enzymes to solid supports. When I started my academic career here at Tufts, I continued working in that area. I started reading about sensors and realized that the field needed organic chemists who understood how to attach things to surfaces. So, I started developing basic methods for attaching molecules to surfaces. For example, we were the first group to use the avidin-biotin method for making sensors back in the mid-1980s.

I hooked up with some people who were developing optical sensors at Lawrence Livermore National Laboratory. They were interested in attaching dyes to optical fibers. I got involved in a startup in the Bay Area that was pursuing blood gas monitoring. We started by taking very thin optical fibers, attaching different dyes to the surface, and then putting them into the bloodstream to get continuous measurements. Through that exposure, I got interested in multiplexed analysis, which led to the idea of creating arrays of sensors. As the desire to make an increasing number of measurements of many species began to develop in the late 1980s, we came up with the idea of creating arrays using imaging optical fibers. We published a paper in Nature on building a multiplexed sensor array using optical fibers. At the time, we prepared these arrays in a very serial way with many steps and many washing procedures, but those days were the early years of the technology.

What is the relationship between what you developed in the late 80s and the platform that Illumina has sold since the late 90s?

Back in the late 1980s and early 1990s, a graduate student named Steven Barnard and I got our hands on some optical imaging fibers. We were putting little spots of polymers on the end of the fibers — very microscopic spots using photo deposition. That idea was published in Nature and was written up in the Wall Street Journal as a new approach to sensing. But it took another five or six years before we invented what is now referred to as the bead array technology that was licensed and commercialized by Illumina.

The idea came about due to one of those serendipitous events in research. One of the students, a postdoc named Paul Pantano, was working on a totally unrelated project, trying to make little points on the ends of fibers for a microscopic imaging application, and every time he did it, he got little wells on the ends of the fibers. We kept wondering, "How do we not make these wells? This is a really awful problem." Eventually, Paul figured out how to make the sharp points. About six months to a year later, I suggested to another student, Karri Michael, that maybe we could fill these wells with beads. Karri had the vision and foresight to conduct that experiment. Karri took some latex beads matched to the size of the wells and suspended [them] in water and applied them to the end of the fiber. After the water evaporated, she observed that the beads settled into the wells. At the time, we didn't know what it was good for. We simply developed the assembly techniques and published a paper in the journal Chemistry of Materials.

[ pagebreak ]

Remember, this was back in the mid-1990s. In those days, researchers were performing low-plex DNA analysis. We were creating arrays on the ends of fibers by the very arduous process of serially spotting polymers. As I was talking with genetics experts, I learned that there was a high probability that people might be interested in measuring a hundred or even a thousand sequences simultaneously. It was one of those special moments in one's career where you remember exactly where you were and all the events leading to a breakthrough idea. The key moment was learning that it may be important to analyze a thousand DNA sequences.

I was working at that time with Columbia University chemistry professor Clark Still on another project involving combinatorial chemistry that used encoded beads to create combinatorial libraries. When I thought about how one could create arrays of thousands of different DNA sequences, it occurred to me that we could take encoded beads, put them in the microwell arrays, use the encoding to identify the bead locations, and thereby simplify array preparation. The whole concept of random arrays came about in a brief flash. I dashed into my laboratory, went up to four students and said, "You, you, you, and you — come into my office." I then proceeded to describe the idea of taking Karri Michael's successful experiment of putting beads into wells, and how by encoding those beads, we could create random arrays. Within a few months, we had reduced the idea to practice and filed all the patent applications. I started talking about the technology with some VC groups and Tufts soon licensed the intellectual property to what is now Illumina.

It's now been 11 years since Illumina was founded. The company is still using random array technology. Some of the products employ the optical fibers we developed. While the format is very similar, they developed some different decoding protocols that enable extremely high levels of multiplexing. Illumina has also moved the random arrays to planar platforms using silicon chips but the basic technology remains the same. Illumina does other things besides arrays now. They are performing next-generation sequencing and have acquired a number of technologies over the years — Solexa, CyVera, and Avantome.

How has the technology developed in the lab since then? Have you further developed that platform or is Illumina supplying you with your research tools?

It's a combination. We have a relationship with the company where occasionally we get some custom arrays. Of course, we have to pay for them. A few years ago, we had some applications in the area of biothreat detection where we developed some new content. Once we fixed the content, we went to Illumina and asked them to make a large number of arrays so we could deliver them to the government. They wanted several hundred arrays that they could test and use. That's an example of where there is no reason for us to devote resources to make arrays. I run an academic laboratory and a company could make them much better than we can.

We also have independent projects where we use similar technology, such as for vapor sensing. One of my graduate students, Matt Aernecke, received his undergraduate degree in forensic science, and he has developed a portable device for sensing different organic compounds at suspicious fires. Instead of having to take samples back to the lab, you can just sniff around and see if there's been lighter fluid, charcoal starter, or other fire accelerants used. We have developed a random array for performing protein-nucleic acid detection. We have also developed an integrated diagnostic system for saliva, where the bead array is integrated into a cassette designed in collaboration with Mike Ramsey's group at the University of North Carolina. So, there's clearly been a diverse set of applications developed for the random arrays. Illumina has a lot of fixed content on their arrays. We tend to do things that are customized and optimized for a particular application. For example, we are working with Don Anderson's group at Woods Hole Oceanographic Institute to develop sensors for harmful algae detection. We are developing arrays that can be used to detect certain algae that cause red tides with the goal of being able to perform an analysis in the field rather than having to take a water sample and bring it back to the lab.

I would say that the technology is by and large similar to what we developed 10 years ago. We just use it for different applications and it's much more integrated. There are other things that have come out of the optical fiber arrays, such as single-molecule detection, which is being commercialized by Quanterix.

You co-founded Quanterix and chair its SAB. Can you give me an update on that company?

We just hired a new CEO. Nick Naclerio, formerly of ParAllele, has been fantastic as the acting CEO since its founding two years ago. David Okrongly, who was previously in the diagnostics group at Siemens and is very familiar with the diagnostics market area, is the new CEO. Nick will remain as executive chairman.

The company has met all its initial milestones in terms of sensitive protein detection at the single-molecule level using the microwell technology that came out of our lab. The goal is to develop very sensitive diagnostics focused on proteins and low-plex assays. The technology involves measuring a small number of analytes at very sensitive levels.

We are now pursuing a number of specific applications in some clinical studies that will give us the opportunity to prove our technology. Things are going well. We'll probably start ramping up a bit more over the next year.

[ pagebreak ]

It seems like you are involved in a number of diverse research projects. How do you manage?

I have a fairly large lab. I like to be involved at the big-picture level. Yesterday, for example, we had a meeting of a subset of the students in the lab. One of our big projects right now is trying to perform single cell analysis, to measure things very sensitively in single cells. Whether that means looking at a large population of cells and determining the differences in protein content amongst the individual cells or looking at an individual cell in a very high resolution fashion by determining the molecular composition using the single-molecule technology we've developed — both approaches get folded into this project. So, I articulate those big picture goals, and then we talk about what are the key experiments that need to be done and the proofs of principle that need to be established for various aspects of the project. But I don't micromanage the group. There were six people at the meeting. After a while, I left the meeting and let the graduate students and postdocs sort it out. It's really up to the individual researchers to decide how they spend their time in the lab. We have a very collaborative lab. People are always talking to each other about various aspects of their projects. I try to bring in people to the lab with new capabilities to help facilitate solving some of these problems.

Because you are developing these technologies, how much do you rely on technology that is produced outside your lab?

We have access to a lot of state-of-the-art instrumentation, here in my lab at Tufts and also the nanoscience facility at Harvard. But those instruments are really for characterization — looking at surfaces and looking at structures. We perform some fabrication outside the lab, such as lithography. On the other hand, we don't want to reinvent the wheel. For example, with our saliva diagnostics project, we want to go from a saliva sample to a result without having to perform any pipetting or other manipulations. We realized that microfluidics was the key. So we hooked up with Ramsey's group at UNC to do all the microfluidics. Our collaboration has been incredibly productive. For the single-cell technology, we are beginning another collaboration. If we see some capability external to the lab that will help jump start our system without having to go buy expensive equipment and learn all sorts of new methods, that's the path of least resistance. I am a chemist by training. Most people in the lab are chemists, physicists, or biologists. They aren't necessarily here to build instruments. We go out, get access to the best people and capabilities, and collaborate in order to bring them to the lab.

You mentioned the single-cell work. Could you give me an overview on what you've achieved there?

All this grew out of our capabilities to perform single-molecule detection. We demonstrated the basic concept using enzymes in solution. We have a continuing effort on the fundamental biochemistry of single enzyme molecules and are looking at their behavior, such as understanding aspects of binding and inhibition — very fundamental work. But we are also interested in developing new analytical capabilities. While we have historically conducted RNA and DNA analysis, we now have the single-molecule technology, and as a consequence, it's opened up all kinds of things we can look at. For example, one would like to be able to correlate messenger RNA content with the amount of protein that's in a cell. The best way to do that is to not amplify either of these species. We are developing techniques for counting the number of messenger RNA molecules and simultaneously counting the number of protein molecules. Again, all made possible by our ability to detect single molecules. To be able to look at multiple cells using the same technology gives you insight into stages of cell development and differentiation and potential defects in metabolism that could eventually lead to disease.

While my lab is not focused on fundamental biology, we are developing technology that will allow us to answer some of those questions. Once we get that technology down, we have some good collaborations in place, such as with Ken Paigen's group at the Jackson Lab. We are interested in applying our technology to big biological problems. We are thinking about the problems and how we can develop technologies that will solve these problems. We are at the very early stages of this project. It will take some breakthroughs and hard work on the part of the students to develop these methods.

If other researchers want to use your technology, should they partner with you, or do you have a path for commercializing it in the future?

It's too early to think about commercialization. One of the things I have learned over the years is that just having a technology or platform or tool is not necessarily the best way to commercialize anything; you really have to have an application. Illumina was an example of a company that was founded on a technology that did not have an application. Frankly, the venture folks thought there were a lot of different paths the technology could take. They thought it could be used in multiplexed analytical chemistry or vapor detection; genetic analysis was just one of the applications where they thought the technology could be used. As it turned out, the folks who decided on the initial application area picked it right. If they had picked something else, it would be hard to believe it would have been as successful as it's been. With Quanterix it was the same kind of thing, but I think everyone recognized the technology would have application in clinical diagnostics.

[ pagebreak ]

Still, it's not clear what the main application of the single-cell work will be; maybe looking at stem cells or defects in a limited tissue sample or being able to do analysis of a biopsy by looking at gross histology followed by a genetic or protein analysis on a cell-by-cell basis. But, in terms of commercializing what we are now working on, it's way too premature. I give talks on entrepreneurship, and one of the things I say is don't start a company until you know it's the right time. I don't mean just raising the money. You have to be very careful — not just about financial resources, but human resources as well. You shouldn't get people excited about joining a company that lasts for a very short amount of time — two or three years or until the money runs out.

What do you spend most of your time doing?

I am involved in a diverse set of projects with a number of different institutions — some academic, some commercial. On the research side, there are several projects. The single-cell and single-molecule work are what I am particularly excited about right now. It's a fairly new area for the group. I deliberately decided to refocus the group on that area, and made a decision to move away from other areas. Education — I really enjoy teaching and I really enjoy mentoring and training the next generation of scientists. This past summer I had 14 undergraduates from all over the country working on projects. We reach out to high school students. For example, we do microarray experiments at high schools in Somerville, Medford, and Malden, and we are trying to expand that program so that young people have an appreciation for the technology. I am also involved with a number of companies but I work with a select and limited number of them to help formulate their strategies. So, my brain gets stimulated by a number of different things.