Name: Gustavo Stolovitzky
Position: Manager, IBM Functional Genomics and Systems Biology group, IBM T. J. Watson Research Center, Yorktown Heights, NY, since 1998
Experience and Education:
Postdoctoral associate, Rockefeller University, 1994-1998
PhD in engineering and applied sciences, Yale University, 1994
MSc in physics, University of Buenos Aires, 1987
Gustavo Stolovitzky heads the functional genomics and systems biology group at IBM Research, where he works on reverse engineering biological circuits, mathematical modeling of biological processes, and new technologies for DNA sequencing.
Last month, the National Human Genome Research Institute, under its "$1,000 Genomes" grant program, awarded him and his team $2.6 million over three years to develop an electrical device, called a DNA transistor, for controlling the translocation of DNA through a nanopore.
A week ago, In Sequence spoke with Stolovitzky about the project, and where it is headed. Below is an edited version of the conversation.
Why is IBM interested in DNA sequencing technology? How did this project come about?
We started working in this field about three years ago. At IBM, we have a strong computational biology team that started in the mid-90s, so we have some expertise in genomic analysis, genome-wide association studies, systems biology, bioinformatics, and a number of other areas of computational biology.
At the same time, we are immersed in a company that is very strong in nanotechnology and silicon technology — that's one of the fortes of IBM Research. About three years ago, I ran into my good colleague Stas Polonsky. He was saying, 'Look, I would like to work on biology and something related to exciting biological research.' I said, 'I'm sure there is room for interaction, there are a lot of things that we can do using nanotechnology and your expertise' — he is a silicon technology guy, and he knows a lot about integrated circuits characterization — 'so maybe what we can do is work on some new way of sequencing.' I, as a practitioner of computational biology, had a clear sense that sequencing was going to be the entry point to a lot of new technologies. It was clear that sequencing was coming, and was coming big.
We decided to immediately think about nanopores, because the notion of an all-electronics technology that has minimum sample preparation and that can measure the identity of the bases based on electronics and not optics was something that seemed within the realm of our expertise within IBM. That’s basically how it started, in a corridor conversation between Stas and myself. Probably no more than a few months after that, we started to write [the] main patent [on a DNA transistor] — we have a number of patents now in different stages of disclosure — and we decided to write our first paper [published in Applied Physics Letters in 2007] and to try to tell our management that this is a nice field to be in. And little by little, we created a group that is growing, in part with the push from the NIH grant, and in part from other potential commercial developments that come down the road. Now we have several papers that have been submitted and we are in the process of publishing.
[ pagebreak ]
Can you describe what a DNA transistor looks like, and what role it will play in nanopore sequencing?
The DNA transistor is a way to integrate silicon technology and different nanotechnological advances. We use little electric fields that we put in electrodes strategically placed within the DNA nanopore to control the electric field inside the pore in such a way that we can trap the DNA as it goes through the pore and release that trap at will. We use that to be able to measure the identity of the bases that are in the nanopore.
Using several nanotechnology steps, we create what we call a sandwich, or a stack of layers, formed by three layers of metal divided by the two layers of a dielectric, or an insulator. These electrodes, which have a width on the order of 3 nanometers, provide contacts to the external world that can be accessed for the experiment. The nanopore is so small that you have to use special drilling tricks. We use an electron beam that drills holes of 2 nanometers and above in width. That hole goes through the stack of metal and dielectrics, and voilà, you have the device. You can watch a video about it here.
The fabrication steps are done in what we call the Microelectronics Research Lab at IBM Research, which is a top-notch facility to do all kinds of trickeries in nanotechnology.
How far have you developed this, and what have you done with it so far?
We have built it, and we are in the process of characterizing and optimizing it. We have already done some measurements of the distortion of the electric field within the pore as sensed by one of the electrodes, and a number of other things that we haven't yet published, so I prefer not to disclose them.
Have you already studied it with DNA inside?
Yes, and we are in the process, over the next few months, of characterizing the force that the electric fields are producing on the DNA, and also the slow-down of the DNA through the pore.
What role do molecular simulations play in your work?
We are using the IBM supercomputer that we have here at the Watson Research Center, which used to be the second-fastest supercomputer at some point, to simulate the processes that take place inside the nanopore, how the DNA experiences the electric field.
When we do molecular simulations, we have all the atoms that occupy a particular space, so the atoms in the DNA, all the water molecules, and all the ions at a particular salt concentration. We then basically use the computer as a microscope into the atomic interactions. We can measure numerically the force that the DNA is experiencing, the time it takes for the DNA to jump from one trapped state to the next trapped state. In each trapped state, we can measure the time as a function of the voltage in the metal layers. We can estimate how much voltage we need in order for our measurement to be precise, long enough. In other words, we are using these as a laboratory that allows us to interpret the experiments by looking at the model of the system. Really looking at the system is impossible from the point of view of optics, but the computer allows us to develop intuition as to what are the processes that are taking place inside the nanopore.
Based on your simulations, how will the DNA behave inside the nanopore?
One thing that we measure is the effective electric field that the DNA is experiencing. DNA has a backbone that has phosphate groups that are ionized in solution. Those negative charges interact with the electric field. The Brownian motion and the ions in the surrounding [area] and the different motions that the DNA is submitted to makes the electric field and the charges interact with an effective force, which is the sum of all the effects that I mentioned. We can measure that net force experienced by the DNA, and one thing we recognize is that the actual force that is experienced by the DNA due to the electric field in the presence of all this molecular randomness and ionic screening is smaller than what we had theoretically calculated. So one thing that we know is that we will have to increase the voltage to trap the DNA. We are using this to optimize the design. At each step, we have an experimental result. We can also look at whether the conventional theory is enough to explain the experimental results or not.
[ pagebreak ]
What do you need to achieve in order to be able to measure the bases at single-base resolution?
You need to make sure that for a sufficiently long time, the DNA molecule is stagnant, so that you can measure, over and over, a particular set of bases that are dwelling in the right place in front of a detector. The detector could be the DNA transistor itself or it could be somewhere else. Within the pore, we could use the electrodes as part of the detector, or we can use slightly removed — but still in the pore — additional electrodes that operate as sensors to identify the nucleoside bases.
What are the challenges you still need to overcome to get to that point?
We are making a first effort to control the translocation of the DNA, and in a second stage, we will concentrate on the sensing part. There are many groups that are working on the problem of sensing the DNA, using nanowires or other nanotechnologies. We are hoping that we will learn from the research these other groups are doing. At the same time, we believe that we have a good solution when it comes to translocation control. Both aspects are necessary.
Are you collaborating with other groups on the DNA transistor project?
We plan to collaborate but we have not established these collaborations. But we are looking forward to collaborate with a number of groups — some groups have solutions for sensing.
Several groups are working on developing DNA nanopore sequencing commercially. What is IBM's intellectual property position in this field?
We are building our portfolio. We are more or less a recent entrant in this field, compared to the earlier patents. Our main patent [on the DNA transistor] is in the works, we have filed it, it's published, and we expect it to issue soon.
What are your plans for commercializing this technology?
We are in discussions with some potential partners. We hope those discussions will pan out eventually, and we will announce that in due course.
How long will it take for this to become a feasible device to sequence DNA?
We hope to have a prototype in something like three years. If the prototype works as expected, probably we need a couple of more years for something a little bit more commercializable.