Where Industry Meets Innovation

  • Contact Us
  • Privacy Policy
  • Copyright
  • Credits
  • sign in Sign In
April 18, 2014Night pic of MIT dome.


Browse News

  • View All
  • ILP News
  • MIT Research News
  • MIT Sloan Management Review
  • Technology Review

ILP Institute Insider

April 7, 2014
News Feature Thumbnail

Mastering the Biological and Engineering Worlds

Rahul Sarpeshkar describes himself as an “amphibian” researcher. It’s a necessary quality in order to do research in the wet and dry worlds that he inhabits and bridges. As head of the Analog Circuits and Biological Systems Group, he designs biological systems in two domains: the wet world of DNA-protein molecular circuits in living cells and the dry world of electronic circuits on supercomputing chips.
Professor Rahul Sarpeshkar
Research Lab of Electronics
Since energy usage is one of Sarpeshkar’s prime focuses, his lab also looks to create ultra low-power medical implants and the renewable energy cells that power them. The ultimate effect of his work and research means improved molecular sensing and synthesis, which could transform implants to aid with hearing, sight, and paralysis, could improve the treatment of systemic diseases, such as cancer and diabetes, and could reengineer living cells to sense, actuate, compute, communicate, and provide energy.

Moving in Two Worlds
The ability to do both wet and dry work at an advanced level is not typical. Sarpeshkar can do it because he made a fundamental discovery that the same thermodynamic physical laws that cause electron flow in transistors also govern molecular flow in chemical reactions. The effect of such a finding is far-reaching. “I can literally take some advanced electronic analog circuits and then map them into DNA-protein circuits, and vice versa,” he says. “It gives us a very unifying way to synthesize and analyze circuits in both the wet and the dry domains.”

The deep connections between electronics and chemistry that Sarpeshkar identified offer possibilities for fundamentally rethinking design methodologies for biotechnology, medicine, and energy. The ability to engineer cells means improved molecular sensing, processing, and synthesis for the pharmaceutical, energy, and food industries. In medicine, there’s the potential for engineering cells to better treat cancer and diabetes, e.g., for designing immune cells that could not only detect a cancer cell but also kill it.

On the supercomputing chip side, Sarpeshkar says that his research works in the opposite direction, by helping to discover problems with cancer and diabetes genes, drug design and cancer treatments through highly parallel computational modeling. His supercomputing chips will also provide a tool to design biological circuits in cells efficiently.

Underlying all of his work is the idea of energy efficiency. One of his main projects has been designing ultra low-power, implantable medical devices, particularly for the deaf, blind, and paralyzed. Sarpeshkar has been working on creating a fuel cell that harvests energy from glucose, metabolizes it, and creates a battery to power these and other medical implants.

As Sarpeshkar says, glucose makes sense as a power source. It’s energy-dense – consider the sustaining effect of one teaspoon of peanut butter. His fuel cell would actually mimic how biology works since biological cells also harvest energy from glucose and use it to make a proton battery that powers their needs. The upsides are that the implant would be antenna-free and battery-free, last the lifetime of the patient and be self-powered without needing any more complicated intervention than eating. “As soon as people power themselves, they power the implant,” he says, adding that the technology could apply to brain implants and pacemakers. “These glucose-powered medical devices could revolutionize the field of medical implants,” Sarpeshkar says.

The Power of Analog Circuits and Analog Computation
There are essential reasons that Sarpeshkar concentrates on analog circuits and biological systems. Analog computation is significantly more efficient than digital computation and it’s the way that biological cells naturally operate. Digital computation uses 1’s and 0’s, on’s and off’s, and Boolean logic to compute. The purely digital strategy is not sufficient for the operation of biological cells to be efficient.

Cells have an incredibly limited amount of energy per unit time that they can metabolize to power themselves, and also a highly limited set of parts, associated with only about 20,000 genes. Cells must be reengineered efficiently or the metabolic burden and high molecular toxicity of the synthetic circuits can become too high and cause them to fail. Therefore, both synthetic and natural circuits need to get more out of less. Since cells cannot afford to use only black-and-white digital signals and logic to function and compute, Sarpeshkar has shown that they must also utilize the power of grayscale analog signals and the power of analog physics and chemistry to compute.

Over the years, his group’s research has shown that analog computation is practical. One of his group members recently engineered a bacterium to compute a square root and other analog functions using only two genetic parts. Others who tried to do the same computation digitally had to use 130 molecular parts. “Analog synthetic biology has the promise of scaling,” he says, adding that it naturally solves problems that have prevented purely digital synthetic biology from scaling.

In another example of the wet and dry worlds coming together, Sarpeshkar has worked on building revolutionary electronic circuits that are inspired by biology. A few years ago, he and researchers in his lab created a radio-frequency chip that could analyze wireless signals over a wide spectrum extremely quickly. This chip mimicked how the ear analyzed the spectrum of sound signals. Currently, he says that he’s investigating circuits inspired by cell biology, which are able to do massively parallel collective analog computing that is insensitive to noise and that can perform reliable computation with unreliable parts.

Speaking Different Languages
For Sarpeshkar, there’s a practicality in seemingly disparate parts collaborating. It applies when working outside of the university setting as well. While there may be different bottom lines, he says that engaging with the private sector grounds his research by bringing in a real-world focus. As an example, he says that small details that might not always be considered important by academia end up being essential when working with industry. Quoting Ludwig Boltzmann, whose laws of physics are the ones that Sarpeshkar exploited to unify biological and electronic circuit design, he says, “That which is the most fundamental is also the most practical.”

And in his lab itself, there’s the interdisciplinary effect from mastering the wet domain of biology and the dry one of engineering. “It’s a synergistic positive feedback loop,” he says. But he does not think that interdisciplinarity automatically manifests by just assembling a team of experts. It is difficult to bring together a biologist and engineer and expect them to speak the same language or to communicate deeply. As Sarpeshkar says, fish don’t naturally interact with land animals; they inhabit different domains. But by bridging the two worlds in one person or intimately in one unified way of thinking, the possibilities increase. “If you’re an amphibian you really can understand both the commonalities and the differences in both fields deeply and work with high creativity and discipline,” he says.

Research News

April 3, 2014

How electrodes charge and discharge

The electrochemical reactions inside the porous electrodes of batteries and fuel cells have been described by theorists, but never measured directly. Now, a team at MIT has figured out a way to measure the fundamental charge transfer rate — finding some significant surprises.

The study found that the Butler-Volmer (BV) equation, usually used to describe reaction rates in electrodes, is inaccurate, especially at higher voltage levels. Instead, a different approach, called Marcus-Hush-Chidsey charge-transfer theory, provides more realistic results — revealing that the limiting step of these reactions is not what had been thought.

The new findings could help engineers design better electrodes to improve batteries’ rates of charging and discharging, and provide a better understanding of other electrochemical processes, such as how to control corrosion. The work is described this week in the journal Nature Communications by MIT postdoc Peng Bai and professor of chemical engineering and mathematics Martin Bazant.

MIT Sloan
Management Review

April 1, 2014

Ethnography in Action at Wells Fargo

Ethnography is described in a recent MIT Sloan Management Review article as “artful in situ investigation into what customers do and feel, and how they talk about what they do and feel.” It’s a disciplined way to try to understand how consumers live, work and play — and how their lives make them more or less receptive to a company’s products and services.

In the article “Stories That Deliver Business Insights,” in the Winter 2014 issue of MIT Sloan Management Review, Julien Cayla, Robin Beers and Eric Arnould give an example of how ethnographic insight helped Wells Fargo Bank develop a behavior-based segmentation that divided retirement-planning approaches of its customers into three categories.

“Only a few years ago, the corporate view of retirement planning at San Francisco-based Wells Fargo Bank tended to focus on dollars and cents — how much an individual needed to invest, by when and for how many years,” write the authors. This segmentation did not account for context such as whether a person was inclined to think about long-term financial goals.

“As part of an ethnographic project commissioned by the bank, researchers had customers walk through a life timeline and recount activities they engaged in that related to retirement planning in each decade of their lives — their 20s, 30s, 40s, 50s and beyond,” write the authors. The stories showed that baby boomers faced “a complex phenomenon of continually negotiated personal travails and marketplace dynamics.”

As a result of what they heard, the Wells Fargo team reworked how they think of customers. The bank developed a behavior-based segmentation that divided retirement approaches into three groups — Reactor, Pooler and Maximizer. These groups represented how people felt about and addressed retirement planning. Reactors are stuck in the now. Poolers tackle financial goals one at a time and tend to be risk averse. Maximizers, the smallest group, think strategically and are willing to make their money work for them. The bank discovered that the category people are in emerges when they are in their 20s or earlier, and is influenced by how many resources they have available and by their general financial savvy.

“This deceptively simple model transformed the way Wells Fargo executives thought about customers’ retirement planning,” write the authors. “The management team grasped that the language of Maximizers (which is closest to the language of bankers) would not resonate with Poolers, who were the largest and probably best-suited target for the company.”

As a result, the bank adjusted its marketing strategy and “designed its retirement planning site to include the various life stages used in the ethnographic research to convey the message ‘we meet you where you are’ and provide relevant, unintimidating guidance — as opposed to producing numbers-dense material filled with endless financial projections.”

In both tone and content, “the overall research result was a more nuanced view of how the interplay between the personal, social and cultural affects people’s vision of retirement planning.”

This article draws from “Stories That Deliver Business Insights,” by Julien Cayla (Nanyang Technological University), Robin Beers (Wells Fargo Bank, N.A.) and Eric Arnould (University of Bath), which appeared in the Winter 2014 issue of MIT Sloan Management Review.