The Genetic Reformation: Rethinking Autonomy and Data Privacy

Topicsfall 2011

Co-hosted by Cherry A. Murray, Dean of the Harvard School of Engineering and Applied Sciences (SEAS), and Leslie Berlowitz, President of the American Academy of Arts and Sciences, two panel discussions examined the “promise and perils” of creating digital repositories for genetic records and considered the policy implications of an individual’s right to access, control, and interpret his or her own genetic data.

The interdisciplinary event, drawing on expertise across the Harvard campus and from around the world, was held in conjunction with a Stated Meeting of the American Academy and regional meetings of the National Academy of Engineering and the Institute of Medicine.

This edition of Topics explores in greater depth some of the issues raised at that “Triple Academies” event.

The feature article (facing page) by Simson L. Garfinkel, Associate Professor at the Naval Postgraduate School, a noted technology writer and former postdoctoral researcher in computer science at SEAS, is intended as an engaging starting point for discussion. (The views expressed are not necessarily those of Harvard or SEAS, but rather are meant to provoke further debate and exploration.)

To read supplementary articles about the issues discussed herein, or to watch the video of the entire “Triple Academies” event, please visit:

http://seas.harvard.edu/topics

Topics | Fall 2011

On April 14, 2011, experts in medical ethics, law, public policy,

research, and entrepreneurship gathered in Cambridge for

a symposium on “Privacy, Autonomy, and Personal Genetic

Information in the Digital Age.”

Roughly nine months before you were born, your biological mother and father wrote a book. They filled that book with their hopes, dreams and plans for your life. They wrote about the adversities you might have to overcome. And they inscribed your family’s secrets—long forgotten infidelities, insanity, and distant cousins who might be monsters.

Now imagine that the book was locked away and lost—only to be found decades later by the trustee of some scientific organization. That trustee may hold the keys to your future, for you are the very person that your parents wrote about. But you’ve also changed—you are much more today than you were when it was written. Do you have a right to decide who reads that book, once it’s found?

And if so, should you read it?

eliza grinnell

Topics | Fall 2011

2 3

Various forms of genetic testing have been available for decades. One of the first widespread testing efforts started in 1969 among Ashkenazi Jews to see if they were carriers of Tay-Sachs, a recessive genetic disease. Because it is recessive, Tay-Sachs has a 25% chance of striking the child of two carriers. But until the testing effort, no one knew who the carriers were. Because the disease is always fatal, the testing effort had but one achievable goal: prevent the conception (or at least the birth) of children who would surely die. The success of this program was one of the first great achievements of genetic testing.

Today genetic testing is widespread. New York state, for example, mandates the screening of newborns for 40 different diseases and disorders. Most of these diseases impact fewer than 1 in 10,000 newborns and can be readily treated with a special diet. Because of the testing, many children are able to lead healthy lives—children who otherwise would have died.

For example, 1 in 19,000 children are born with phenylketonuria (PKU), a disease characterized by an inability to metabolize phenylalanine, a commonly occurring amino acid. People with PKU who “diet for life” (by avoiding milk, eggs, the artificial sweetener aspartame, and other foods) are able to lead normal lives. Those who don’t, suffer delayed development, mental retardation, and a variety of other problems. Another victory for large-scale genetic testing.

Genetic testing also gave rise to a new professional class—the genetic counselor. Like the priests of old, these people were trained in the intricacies of an unfamiliar language—although this language was

the As, Cs, Gs and Ts of the genetic code, rather than the tempus nascendi, et tempus moriendi [“a time to be born and a time to die”—Ecclesiastes 3:2] of the Latin Vulgate Bible. But like priests, genetic counselors were intermediaries, standing between the laity and a higher authority. And they were needed, because until the 1990s, most Americans not only lacked the ability to interpret their test results; they didn’t even have legal access to their own medical records.

Now, for the first time in history, anyone on the planet who has a few hundred dollars (and is willing to spit into a tube for 15 minutes) can get vast amounts of genetic information with no intermediary whatsoever. In a few years, you’ll be able to get your entire genomic sequence for less than $1,000. (You can order it today for $4,995 from Knome, Inc., a life sciences company in Cambridge, Mass.) Last year, an advisory panel told the U.S. Department of Defense that it needed to start planning for the advent of the “$100 genome”—and with it the possibility that American soldiers might be covertly tested by the enemy.

Just as Bibles translated into the vernacular helped power the Protestant Reformation, direct-to-consumer (DTC) genetic testing is opening the door to a genetic reformation. That reformation will funda-mentally change our notions of ourselves, our place in the world, and our human potential. And anyone who takes the plunge will find that this genetic data brings them into a new world—one in which tradi-tional authorities have less influence and individuals have less privacy and greater risk. And yet, the actual scientific payoff is still largely unknown.

Several websites, such as deCODEme.com and Navigenics.com, offer a variety of genetic tests directly to the consumer for between $500 and $1,500. But the poster child for the genetics reformation is unquestionably 23andMe.com, a Google-backed Silicon Valley start-up that offers broad-spectrum genetic testing for about $100 (provided you sign up for a $5/month monitoring service).

Signing up is easy. Just go to a website; accept the frightening consent statement; type in a credit card number; and a few days later 23andMe’s DNA collection kit will appear in your mailbox. The “kit” is really nothing more than a tube with a fancy lock and some preservatives. Avoid eating for 30 minutes; fill the tube with spit; snap on the cap; and send it back. A few weeks later you’ll be able to browse your risk factors for more than 90 diseases and traits on the company’s website.

But genetic analysis is only the beginning of what 23andMe does with your data. Recall that each of your parents contributed half of the words in your genetic book. This means that if you and your long-lost sister (or half-sister) both sign up for 23andMe, the company’s computer can match the two of you by the fact that 50% of your genetic material is in common. And here’s where things get sticky. 23andMe can also determine to a high probability

that you and that woman across town have the same father but different mothers; it will even let you contact each other through the company’s website.

You didn’t know you had a half-sister? Oh my! That’s why 23andMe’s consent statement reads, in part, “You may learn information about yourself that you do not anticipate,” and “Once you obtain your genetic information, the knowledge is irrevocable.”

For investors, another significant part of 23andMe is likely to be the company’s research arm—23andWe. Human genetics is vastly more complicated than the simple Mendelian genetics that most of us learned in high school. Disease susceptibility, drug response—even “simple” things like hair color and curl—actually result from the interaction of multiple genes along with environmental challenges and other factors that modern biology is only just beginning to understand. So in addition to testing the genotype of each subscriber, 23andMe invites them to participate in research studies by answering detailed questionnaires about their phenotype—that is, their medical history, traits, and morphology, such as the shape of their noses and the character of their ear wax. All of these data are mined in an effort to draw correlations between genetic sequence, medical conditions, and medical outcomes.

The Personal Genetics Revolution, Right In Your Web Browser

Just as Bibles translated into the vernacular helped power the Protestant Reformation, direct-to-consumer genetic testing is opening the door to a genetic reformation.

cour

tesy

of

sim

son

garf

inke

l

To read a personal account of the author’s experience with genetic testing,

visit http://seas.harvard.edu/topics.

4

Topics | Fall 2011

Unlike the 40 tests mandated by the state of New York, 23andMe isn’t performing traditional genetic tests. When children in New York are tested for PKU, they are tested to see if an enzyme called phenylalanine hydroxylase is present and properly functioning—in other words, that the individual’s genes are correctly producing that enzyme. 23andMe doesn’t assess the presence or absence of given enzymes or proteins. Instead, it screens for what are called single nucleotide polymorphisms —SNPs (pronounced “snips”) in the language of the new biology. A SNP is a variant spelling in the genome written by our biological parents.

If you think of each person who’s alive today as having a book in humanity’s genetic library, it turns out that those 7 billion books are remarkably similar. In the vast majority of cases, these books have 46 chapters, representing 23 chromosomes from each parent. Each book is thought to have 20,000 to 25,000 pages—one page for each protein-coding gene. Though there are roughly 3 billion base pairs in total—3 billion As, Cs, Gs, and Ts—most of these don’t code for genes, but seem to be associated with some form of regulation. Yet, in all of this writing that’s so important in determining our destiny, there

are only about 10 million places where one person’s “T” might be another person’s “G,” or where another person might have a few extra Ts—as if his or her mother or father momentarily stuttered. And according to the 23andMe website, the company now measures roughly 1 million of these 10 million SNPs from the 23 chromosomes, as well as a few thousand SNPs from the mitochondrial DNA.

Although reading the SNPs is a highly precise technique for measuring a person’s genetic profile, for many genetic diseases it is far more accurate to simply measure the presence or absence of a functioning enzyme. That’s because, in the case of PKU, there are literally hundreds of different genetic mutations that might cause a child to have an absent or poorly functioning phenylalanine hydroxylase enzyme. While some of these mutations are known and reported in the medical literature, others aren’t.

For the cases where the mutation is known, the 23andMe website tells subscribers if their SNPs match the literature. The website will even cite the study, allowing consumers-turned-scientists to examine the literature for themselves.

For cases in which the link between the genotype (the specifics of the genetic plan) and phenotype (the expressed, physical characteristics) is unknown, 23andMe hopes that those surveys will help scientists to draw correlations between various SNPs and the prevalence of various diseases. For example, if 75% of 23andMe’s subscribers who have a “T” in a particular position tend to get some disease by age 30, and most people who have an “A” there do not, then that “T” might be associated with a 75% chance of contracting the disease. Then again, it might not.

To put this into actual genetic terms, consider leucine-rich repeat kinase 2, an enzyme coded by the LRRK2 gene, one of the 1,370 genes on chromosome 12. A specific mutation in this gene called rs34637584(A) is associated with a significantly higher chance of contracting Parkinson’s disease. Years ago, finding out what such a mutation meant would have required a master’s degree and access to a medical library. Today you can type the mutation into Google and be directed to a page on SNPedia, an open-source wiki devoted to collecting such information and making it public. The wiki says, in part:

One copy of a[n] rs34637584(A) allele is sufficient to greatly increase one’s risk for Parkinson’s disease. … Overall, the risk of Parkinson’s disease for a person who inherits a[n] rs34637584(A) allele is 28% at age 59, 51% at 69, and 74% at 79, according to the International LRRK2 Consortium.

Unfortunately, the wiki misstates the evidence.

As all students learn during their first statistics course, correlation is not causation! The rs34637584(A) allele may be more likely to be present in people who have Parkinson’s disease, but we don’t know if it is the cause—the SNP might work in concert with another gene, or with an environmental agent, or it may be an innocent genetic bystander.

It’s these uncertainties, in part, that have caused organizations such as the American Medical Association, the National Society of Genetic Counselors, and the American Society of Human Genetics to call for significant regulation of DTC companies. In February 2011, the American Medical Association’s executive vice president, Dr. Michael D. Maves, wrote to the Food and Drug Administration, urging that tests “with the highest risk of harming consumers if misinterpreted have the strictest regulatory requirements,” and recommending that companies like deCODE Genetics and 23andMe be legally required to report these test results to a customer’s physician or genetic counselor, and not directly to the consumer.

Indeed, the state of New York already prohibits companies from offering direct-to-consumer genetic tests. As a result, when someone in Manhattan wants to be tested by 23andMe, the company requires that the specimen be mailed from outside New York—for example, by taking a 10-minute subway ride to New Jersey and dropping the package into a Hoboken mailbox.

Doctors and genetic counselors who want legislative prohibitions on DTC testing are clearly acting in their own economic interest: each consumer who bypasses today’s inefficient healthcare system and goes directly to these companies is saving hundreds, if not thousands, of dollars. Wiki-based counseling is free.

On the other hand, it’s easy to take the genetic priesthood’s claims at face value: this is powerful information and easily misinterpreted. There are documented cases of people committing suicide after learning that they were carriers for Huntington’s disease—and those people received counseling in a clinical environment. We have no idea how much damage might be done in the coming years by the casual release of such sensitive medical information.

DTC’s Organized Opposition

5

patrick landmann / pho

to researchers, inc.

Advances in bioinformatics—the application of computer science to biology and medicine—have been crucial for the field of genomics. Fast and accurate assembly of complete genomes would not be possible without sophisticated sequencing algorithms, modeling techniques, and (as shown above) data visualization tools. In this computer readout, each color corresponds to a nucleotide (A, C, T, or G) detected in a genomic sequence.

Topics | Fall 2011

7

Another danger is that this information might not be adequately protected. 23andMe allows its customers to download their entire genetic data set: other websites invite you to upload it for a third-party analysis. The problem here, of course, is that unlike a stolen credit card number, a genome can’t be changed if it is inadvertently given to criminals. Is that a risk? We just don’t know.

There’s another potential problem with DTC genetic testing: is it accurate?

“We are concerned about analytic validity,” says Dr. Michele Caggana, Section Head for Genetic Testing for the New York State Clinical Laboratory Reference System. “If you order the test 10 times, do you get the same results 10 times?”

The Government Accountability Office (GAO), the watchdog agency of the U.S. Congress, has twice reviewed DTC testing firms and found troubling inconsistencies, reporting in 2006 that the firms made “medically unproven disease predictions.” A 2010 GAO report—“Direct-To-Consumer Genetic Tests: Misleading Test Results Are Further Complicated by Deceptive Marketing and Other Questionable Practices”—was even more damning. It constituted an outright attack on the industry, accusing it of inconsistent test results, incorrect information delivered by telephone consultants, and the use of genetic information to scare customers into purchasing expensive vitamin supplements.

The GAO did not release the names of the com-panies that it investigated, but it did refer them to the Food and Drug Administration and the Federal Trade Commission “for appropriate action.”

In a high-profile June 2010 incident, 23andMe mixed up the samples in a 96-well plate and sent incorrect DNA results to 96 of its customers. Whoops! One family, which had tested parents as well as children, was taken aback—they thought that their son might have been accidentally swapped at the hospital when he was born (apparently it had happened a month before at the same hospital to another pair of babies). Another company, deCODE Genetics, had a similar problem in August 2009.

For many contemplating DTC tests, issues of cost and accuracy are less important than the potential damage that might come from taking the test—not just the way that genetic information might change one’s sense of self, but the real potential for genetic discrimination.

There’s a long history of using genetics and pseudo-genetics to justify discrimination against individuals and racial groups—and you don’t need to go back to the Second World War for examples. Since the 1980s, the Council for Responsible Genetics has documented more than 500 cases in which apparently healthy individuals have been “barred from employment or lost their health and life insurance based on an apparent or perceived genetic abnormality,” according to CRG’s project on Genetic Testing, Privacy, and Discrimination. Fears of genetic discrimination were also taken to the big screen in the 1997 movie Gattaca.

In 2008, Congress passed the Genetic Information Nondiscrimination Act (GINA), which prohibits the use of genetic information for determining health insurance rates or employment. But GINA does allow genetic tests to be used for setting rates on long-term care insurance and life insurance. This means that women who are tested and found to carry harmful mutations of the BRCA1 or BRCA2 genes—that is, women who have more than a 50% chance of developing breast or ovarian cancer—can’t be denied a job or health insurance, but they can be denied life insurance.

Is genetic discrimination a compelling risk? Dr. Philip Reilly, who spent years caring for institutionalized individuals with genetic disorders and now, at Third Rock Ventures in Boston, invests in companies that are trying to treat them, insists that it’s not.

“We have a 40-year history of gathering, storing, and protecting” genetic information, Reilly says. “There is virtually no evidence that anyone has suffered an economic harm from newborn screening. It’s appropriate to think about the [potential for abuse], but it’s outweighed by the benefits.”

The trouble with this argument is that newborns haven’t been screened for their risk of contracting Parkinson’s or Alzheimer’s diseases later in life— two diseases that have profound financial impact on those offering life or long-term care insurance.

“To what extent can one actually know the conse-quence of releasing that data?” asks Latanya Sweeney (A.L.B. ’95), a Visiting Professor of Computer Science at Harvard’s School of Engineering and Applied Sciences (SEAS). Sweeney contends that we simply don’t know the ways that this information could be abused.

Even if there are no direct harms, many feel that it is a violation of personal privacy to release even anony-mous genetic information without consent. And that’s exactly what happened in Texas between 2002 and 2009, when 8,350 of the 5.3 million samples collected during the course of its newborn screening program were released to 27 separate research pro-grams by scientists around the United States.

In March 2009, shortly after news of the medical research was made public, Geoffrey Courtney of San Antonio and four other parents filed suit against the Texas Health Department and Texas A&M Univer-sity, alleging that the state’s retention of the blood spots and their use in research and federal investiga-tions constituted unlawful search and seizure and violated their privacy rights. The suit was settled out of court, but in response, the Texas legislature passed a law specifically authorizing this use of the

Risky Business

Unlike a stolen credit card number, a genome can’t be changed if it is inadvertently given to criminals. Is that a risk? We just don’t know.

Jonathan Zittrain (M.P.A. ’95, J.D. ’95), Professor of Computer Science and Law at SEAS and Harvard Law School, submitted a saliva sample to a personal genetic testing service to see what it was all about. Now, when he logs into 23andMe.com, the website informs him of his risks and traits, including his earwax type—“wet.”

cour

tesy

of

jona

than

zit

trai

n

6

8

blood spots—provided that the parents were first allowed to opt out of the collection by signing a form. In early 2010, Texas incinerated the 5.3 million blood spots that had been collected prior to the passage of the law.

But there are problems even with anonymous genetic samples. When such “de-identified” samples are released for research, the anonymization is really just a legal fiction. Like a picture of a face or a fingerprint, genes carry information that is hugely personal. With enough ancillary information—for example, a large database that happens to have DNA SNPs from other family members—re-identification is quite possible.

“It might be possible 10 years from now to actu-ally anonymize genetic data, but it’s not possible now because we don’t know what it is we need,” says Sweeney, who has worked on techniques for re-identifying medical records and other kinds of information for more than decade. “If I need all of it, then I can’t really de-identify it, because it’s you.”

Another thing that can’t be de-identified is famil-ial relationships. With a database of SNPs from thousands of people, it’s now fairly straightforward to identify who’s related to whom. And because the DNA molecule is stable over a long period of time—you can recover DNA from corpses that are thousands of years old—blood or body specimens from people who died in the 1950s could easily be used to learn sensitive information about people living today. This creates a paradox under U.S. law, since the dead legally have no privacy rights.

The ready availability of genetic information, made possible by direct-to-consumer genetic tests, thus creates fundamental challenges to our notions of privacy, autonomy, and consent. Given the shared nature of genetic information, it may be a funda-mental misconception to view this data as “private.” Indeed, Dr. George M. Church (Ph.D. ’84), Professor of Genetics at Harvard Medical School and Director of the Center for Computational Genetics, has published a document arguing that there are so many ways that genomic confidentiality might be compromised, that “guarantees of genome anonymity” are simply unrealistic.

Anonymization is really just a legal fiction. Like a picture of a face or a fingerprint, genes carry information that is hugely personal.

There’s a lot we don’t know about the impact of personal genetic information on individuals, families, and society as a whole. Would learning that you have an 80% chance of dying from heart disease by age 40 lead you to embrace a healthier lifestyle in the hope of beating the odds—or simply justify continued pigging out on those high-fat foods, in the belief that you can’t change your genetic destiny? We just don’t know.

Likewise, we don’t know how many complete genomes are needed to make fundamental discoveries. The LRRK2 Parkinson’s study men-tioned in the SNPedia entry was based on a study of just 1,045 people from 133 families—and for those people, the study confined itself to the LRRK2 gene. Perhaps better science would have been possible if more people had been studied, and if the entire genome for those symptomatic individuals had been made available.

Questions like these are at the heart of the Personal Genome Project, a multi-year research effort headed by Prof. Church at Harvard Medical School. This federally funded project is seeking volunteers who will consent to having their entire genome sequenced and made freely available on the Internet with the goal of aiding scientific discovery. But the project also hopes to study the volunteers apart from their genomes, exploring the impact of genetic information and education on them and their families.

Ultimately, the PGP seeks to collect and publish the genomes for 100,000 individuals. The first 10 individuals are also sharing their detailed medical records and other highly personal information. These so-called “PGP-10” include Church; venture capitalist Esther Dyson; the CEOs of several genomic-based healthcare organizations; and research scientists from Harvard and Duke.

For the CEOs and scientists involved there was a clear benefit in making their genomes available: they hope to profit from the availability of scientific research data. But what about the other 99,990 people that Church wants to recruit? For those, the project hopes to attract individuals with a combina-tion of personal curiosity and “genetic altruism” —people who feel that, by sharing their genome, they can help make the world a better place.

Just as the publishing and discussion of Martin Luther’s 95 Theses powered the Protestant Reformation, it’s almost certain that the growing availability of genetic information from DTC genetic testing and online genomic sharing will result in profound changes, producing a genetic reformation. Our challenge as a society is to guide this reformation in such a way that it maximizes the benefits of the information while minimizing the potential for harm.

Making Personal Genomes Public

9

volker steger / pho

to researchers, inc

Dr. George Church, Ph.D. ’84, Professor of Genetics at Harvard Medical School, created the Personal Genome Project in 2005 with the hope of gathering 100,000 individuals’ genomic sequences and medical histories for scientific research. Almost 30 years ago, while earn-ing his Ph.D. at Harvard in biochemistry and molecular biology, he helped to develop the first direct genomic sequencing method. Today, he combines genomics with epigenetics and developmental biology to study stem cells.

cour

tesy

of

flic

kr u

ser

dan

bucz

ynsk

i

To read more about Latanya Sweeney’s work in data anonymity, most recently involving prescription drug records, visit http://seas.harvard.edu/topics.

Topics | Fall 2011

11

Around Oxford Street

Hands onUndergraduates created some dazzling projects this year.

In ES 51, Computer-Aided Machine Design, students built Tootsie Roll catapults and converted cordless electric screwdrivers into remote-control cars.

The Science & Cooking Fair showed off crazy culinary creations like glow-in-the-dark gummy bears and Parmesan noodles.

The CS 50 Fair, Hack Harvard, and the I3 Challenge produced hundreds of new websites and apps, including Hollre, Aid Aide, and Newsle (recently profiled in Tech Crunch, The Huffington Post, The Chronicle of Higher Education, and on the BBC).

In January, undergrads designed a disaster relief tent that was supported by giant helium balloons; and in April, students in ES 96, Engineering Design Seminar, became masters of a geothermal heating and cooling system on campus—studying it literally in depth.

Computational science expandsBeginning this fall, SEAS will offer a graduate secondary field in Computational Science and Engineering. The program, which complements the Institute for Applied Computational Science at SEAS, is designed to equip students with rigorous computational methods for approaching scientific questions across disciplines.

Topics will include modeling and simulation of complex systems; parallel programming and collaborative software development; and methods for organizing, exploring, visualizing, processing and analyzing very large data sets.

Enhancing the learning environmentSeveral new staff hires enhance the design curriculum and bolster our existing resources for undergraduate advising this fall.

Avinash Uttamchandani and Joe Zinter, our new preceptors in design-based instruction, will provide advice and support in the undergraduate hands-on design courses.

Sujata Bhatia, M.D., Ph.D., joins Margo Levine, Ph.D., as an assistant director of undergraduate studies. Bhatia’s and Levine’s activities include teaching, advising under-graduates, and building a cohesive student community.

All Hands Meeting—onward and upwardDean Cherry A. Murray addressed the entire SEAS community at an All Hands Meeting in April. Quick recap: The student population at SEAS continues to grow, and the percentage of Computer Science concentrators who are female has jumped from 13% to 40%. Student innovation, faculty spin-off companies, and alumni business ventures are expanding the influence of SEAS worldwide; and the undergraduate teaching labs are supporting a renewed emphasis on design and hands-on projects in the curriculum.

To watch the video of April’s All Hands Meeting, visit http://seas.harvard.edu/topics

Science & Cooking public lectures returnInspired by one of the most talked-about Harvard College courses in recent history, “Science & Cooking: From Haute Cuisine to the Science of Soft Matter,” the Science & Cooking public lecture series has returned, with weekly lectures throughout the Fall 2011 semester. Members of the general public are once again invited to attend talks by world-class chefs and eminent food experts, including Grant Achatz, José Andrés, Ferran Adrià, Harold McGee, and Dan Barber.

The undergraduate Science & Cooking course will continue for the next 5 years.

To view the full schedule of public lectures, visit http://seas.harvard.edu/cooking

Community Highlights

What do we mean by “privacy”?

Latanya Sweeney (A.L.B. ’95), Visiting Professor of Computer Science at SEAS, likens this question to the picture of the elephant surrounded by blindfolded researchers, each of whom is touching a different part of the animal and wondering what it is.

To some, privacy means liberty: the right not to share; to be free from discrimination; to have individual auton-omy. To others, it means treating information as property that can be owned, protected, and traded fairly and with informed consent. And to others, the expectation of total privacy is just an obstacle to progress and enterprise in a world where, like it or not, information is going public.

Research and thought leadership at Harvard spans all of these attitudes, as well as the grey areas between them.

Harry Lewis (A.B. ’68, A.M. ’73, Ph.D. ’74) warns that ignorance might be the greatest threat to privacy today, given the amount of data already collected by cell phone towers, ATMs, websites, health records, the police, the courts, credit card companies, pharmacies, grocery stores, and so on.

Lewis is the Gordon McKay Professor of Computer Science at SEAS, a Harvard College Professor, and a co-author of Blown to Bits: Your Life, Liberty, and Happiness After the Digital Explosion (2008).

“Every day we encounter unexpected consequences of data flows that could not have happened a few years ago,” Lewis writes, with co-authors Hal Abelson and Ken Ledeen.

“The digital explosion is creating both opportunities and risks. Many of both will be gone in a decade, settled one way or another. Governments, corporations, and other authorities are taking advantage of the chaos, and most of us don’t even see it happening.”

Despite recent shifts in public attitudes about what is appropriate to share publicly—consider the sheer amount of personal information posted voluntarily on Facebook —some types of data (financial, medical) are simply too risky to share and need special protection.

That’s where people like Salil Vadhan (A.B. ’95) come in. Vadhan is the Vicky Joseph Professor of Computer Science and Applied Mathematics and director of the Center for Research on Computation and Society (CRCS) at SEAS.

An expert in cryptography, Vadhan researches topics in computational complexity theory that help protect sensitive data from malicious attacks.

But who decides what should be protected and what should be shared?

Jonathan Zittrain (M.P.A. ’95, J.D. ’95) studies the future of the Internet, including the forces and institutions shaping it.

He co-founded the Berkman Center for Internet & Society at Harvard and holds joint appointments at SEAS, Harvard Law School, and Harvard Kennedy School as Professor of Computer Science and Professor of Law. He advises the Federal Communications Commission and the National Security Agency, and serves on the boards of the Internet Society and the Electronic Frontier Foundation.

Zittrain worries that while we have made progress on traditional privacy problems stemming from corporate and government intrusion, we are now entering the uncharted waters of peer-to-peer privacy breaches.

For example, he says, “It’s becoming easy to signal to the world’s tourists—which is to say, anyone with a cell phone camera—that you’re willing to pay a bit for photos or video of a particular person or place that happens to be nearby.”

“We can achieve (and perhaps regret) ubiquitous surveillance without Big Brother’s involvement at all.”

Experts at SEAS, CRCS, and the Berkman Center generally agree that policy and ethics risk falling far behind the pace of technological change, and that’s very much an ongoing discussion at Harvard.

Perhaps more worryingly, though, public awareness lags too. Helping to spark such awareness—expert as well as public—was one aim of April’s “Triple Academies” symposium.

Rethinking Privacy at Harvard

10

eliz

a gr

inne

ll

Left to right: Latanya Sweeney, Harry Lewis, Salil Vadhan, and Jonathan Zittrain. (Photos by Caroline Perry and Eliza Grinnell.)

14 1312

Research Briefs

Brain injury after the IEDBioengineers identify the cellular mechanisms of traumatic brain injury, offering new hope for the treatment of veterans wounded by explosions.

Who: Kevin Kit Parker, Professor of Bioengineering at SEAS, and a team of interdisciplinary researchers in the Disease Biophysics Group.

How it works: Using new tissue engineering methods designed by undergraduate and graduate students at SEAS, Parker and his team have explained how the jarring force of an explosion disrupts structural networks and signaling pathways in the brain.

What’s next: Membrane proteins called integrins appear to translate external forces into a damaging signal cascade inside the neurons and vascular cells. Future research may develop a therapeutic treatment that inhibits specific proteins within that signal path to prevent long-term brain injury.

Fuel cell breakthroughMaterials scientists demonstrate the first macro-scale, thin-film, solid-oxide fuel cell (SOFC), showing that the technology can be scaled for practical clean-energy applications, such as transportation and electronics.

Who: Shriram Ramanathan, Associate Professor of Materials Science at SEAS, postdoctoral fellow Bo-Kuai Lai, and Masaru Tsuchiya ’09 (Ph.D.), who now works at SiEnergy Systems.

How it works: SOFCs create electrical energy via an electrochemical reaction that takes place across an ultra-thin membrane. Ramanathan’s team scaled the system up by a factor of 100, fortifying the fragile membrane with a nanoscale metallic grid.

What’s next: While SOFCs have previously worked at the micro-scale, this is the first time any research group has overcome the structural challenges of scaling the technolo-gy up to a practical size with a proportionally higher power output. The next steps are to experiment with methane as a fuel source, rather than hydrogen, and to lower the cost of the materials.

Detecting land mines—there’s an app for thatNew smartphone-aided technology simplifies the dangerous task of finding and identifying land mines left over from past wars.

Who: Krzysztof Gajos, Assistant Professor of Computer Science at SEAS, researcher Lahiru Jayatilaka, and collaborators at Carnegie Mellon University and MIT.

How it works: The app collects audio signals from a metal detector and builds up an image of the buried object’s outline—one red dot for each beep—so users can quickly differentiate a scrap of junk metal from an actual land mine.

What’s next: Using this tool, the search for old land mines will be more efficient, as less time will be wasted on investigating false alarms. Recognizing buried mines with just a metal detector requires a great deal of training and experience. With this app, it will be quicker and easier to clear the hazards and reclaim the land.

Tut, tutThe walls of the Egyptian tomb of Tutankhamen are cov-ered in mysterious brown spots that appear to be metabolic stains left by microbes. Analysis suggests that the painted walls were not dry when the tomb was sealed, implying that the young king’s burial was most likely a rush job.

Who: Ralph Mitchell, Gordon McKay Research Professor of Applied Biology at SEAS, with research assistant Alice DeAraujo and postdoctoral fellow Archana Vasnathakumar.

How it works: Culturing, DNA analysis, and chemistry have revealed that the culprit was most likely fungal —perhaps even Penicillium—and that it has long been dead. Mitchell’s work suggests that the moist walls and the organic matter (from the corpse and the ritual food and incense) provided a bountiful environment for the ancient microbes, until the tomb dried out.

What’s next: Conservators must decide whether to leave the spots intact, due to their historical nature, or to clean the walls. Meanwhile, Mitchell’s lab is investigating possible methods of removing the stains without damaging the paint or the plaster.

Hidden message identifies unknown liquidsPortable, power-free device changes color in the presence of certain liquids, even distinguishing among varying concentrations of the same substance.

Who: Professors Joanna Aizenberg and Marko Loncar, with graduate students Ian Burgess, Lidiya Mishchenko, and Mathias Kolle and research appointee Benjamin Hatton.

How it works: The nanostructured chip contains a network of pores with finely tuned optical and chemical properties. A liquid can flow through the pores only if its surface tension is precisely correct. When the surface gets wet, it changes color, indicating that a particular liquid is present.

What’s next: The technology may find applications in quality control, security, identifying chemical spills—even detecting the presence of methanol in bootleg liquor.

courtesy o

f the u.s. department o

f defense

courtesy o

f shriram ram

anathan

courtesy o

f ian burgessAround Oxford Street

Around Oxford Street

14 15

New Faculty Hires

Former SEAS Dean Venkatesh “Venky” Narayanamurti has been appointed foreign secretary of the National Academy of Engineering.

Two faculty members have received tenure: Kevin Kit Parker (Professor of Bioengineering) and Todd Zickler (Professor of Electrical Engineering).

(Right) Joanna Aizenberg, the Amy Smith Berylson Professor of Materials Science, has been named a new director of the Kavli Institute for Bionano Science and Technology at Harvard. The Radcliffe Institute at Harvard has also appointed her a Director of Academic Ventures for its Science Program.

David Malan ’94, S.M. ’04, Ph.D. ’07, instructor for the popular introductory computer science course CS 50, has been appointed Senior Lecturer for Computer Science. He will also take on the role of Director of Educational Technology, helping SEAS and the College to expand the

use of pedagogical technology in the classroom. In the near future, with Malan’s help, we hope to make all SEAS courses available online.

Conor Walsh, Assistant Professor of Mechanical and Biomedical EngineeringWalsh enjoys teaching engineering design through hands-on courses. His research involves developing robotic tools for rehabilitation and surgical assistance, as well as other innovative medical devices.

David Keith, Gordon McKay Professor of Applied Physics and Professor of Public PolicyKeith is an expert on energy technology, climate science, and policy. He will hold a joint appointment with the Harvard Kennedy School.

Leslie Valiant, the T. Jefferson Coolidge Professor of Computer Science and Applied Mathematics, won the 2010 ACM A. M. Turing Award, the so-called “Nobel Prize in Computing.” The award carries a $250,000 prize.

(Above) Roger W. Brockett, the An Wang Professor of Electrical Engineering and Computer Science, was honored with the McDonald Mentoring Award, having advised more than 60 graduate students over the course of his career.

Michael Brenner, Glover Professor of Applied Mathematics and Applied Physics, was awarded the George Ledlie Prize by the President and Fellows of Harvard College. The prize is awarded no more than once every two years to someone affiliated with the University who “since the last awarding of said prize has by research, discovery, or otherwise made the most valuable contribution to science, or in any way for the benefit of mankind.”

Vahid Tarokh, Perkins Professor of Applied Mathematics and Vinton Hayes Senior Research Fellow, received a Guggenheim Fellowship.

(Below) Alice Chen, Ph.D. ’11, won the 2011 Lemelson-MIT student prize. A SEAS graduate of the Harvard-MIT Division of Health Sciences & Technology, Chen created a mouse with a tissue-engineered human liver that can be used in drug trials.

Steven C. Wofsy, Abbott Lawrence Rotch Professor of Atmospheric and Environmental Science, was elected to the National Academy of Sciences.

A project to use dirt-powered batteries to charge cell phones in Africa won a $100,000 grant from The Bill & Melinda Gates Foundation. Aviva Presser Aiden, Ph.D. ’09, and colleagues will help to develop a microbial fuel cell-based charger that could be readily and cheaply assembled out of basic components to increase access to health care via mobile applications.

Debra Auguste, Assistant Professor of Biomedical Engineering, and Stephen Chong, Assistant Professor of Computer Science, both won the National Science Foundation’s Faculty Early Career Development (CAREER) Award.

Crimson wide receiver Baltazar Zavala ’11 (engineering sciences & neurobiology) was chosen as a Rhodes Scholar, receiving the news on the field after November’s victorious Harvard-Yale football game.

Alumni in the News

Faculty NewsSelect Awards

Shiv Gaglani ’10 (Biomedical Sciences and Engineering) and four of his former classmates published a book, Success with Science, aimed at encouraging high school students to pursue hands-on research in science and engineering.

Computer scientist Kim Hazelwood, Ph.D. ’04, was named to Technology Review’s TR35.

Environmental engineer Martha Heitzmann, Ph.D. ’97 has been appointed Senior Executive Vice President of Research and Innovation at the Areva Group, a French energy company that specializes in nuclear power. Read a Q&A with Martha Heitzmann online at http://seas.harvard.edu/topics

Pasquale Pat Romano ’87 was named president and CEO of Coulomb Technologies, a company that creates the infrastructure necessary for charging electric vehicles.

Chris Capossela ’91 (Computer Science and Economics) became the Senior Vice President of Microsoft’s Consumer Channels and Central Marketing Group.

Applied math alum and Futurama writer Ken Keeler ’83, Ph.D. ’90 won a 2011 Writer’s Guild Award.

Dennis M. Ritchie ’63, A.M. ’65, Ph.D. ’68, who studied applied mathematics as a graduate student, was awarded the Japan Prize in January, along with Ken Thompson. Ritchie and Thompson developed the UNIX computer operating system in 1969 when they were both researchers at Bell Labs.

Ryan Adams, Assistant Professor of Computer ScienceAdams is an expert in machine learning and computational statistics.

Eddie Kohler, Assistant Professor of Computer ScienceKohler’s research explores systems, networks, programming languages, and software engineering.

eliz

a gr

inne

ll

stephanie mitchell / harvard new

s office

eliz

a gr

inne

llco

urte

sy o

f al

ice

chen

eliz

a gr

inne

llel

iza

grin

nell

eliz

a gr

inne

ll

Imagine if someone hacked you.

It’s not that far-fetched. A clever techno-thief might find an easy, undetectable means to capture and sequence your DNA. Or, if you already have your genetic information online, a burglar could do the hacking the old-fashioned way.

As Harry Lewis warned in his book Blown to Bits, almost every aspect of our lives is stored somewhere as data. Our digital footprint now extends even to our DNA.

Lewis wrote, “The digital explosion is changing the world as much as printing once did—and some of the changes are catching us unaware, blowing to bits our assumptions about the way the world works.”

In this issue of Topics, writer Simson Garfinkel reports that interpreting the human genome is more like reading a palm than a map. Garfinkel, a former fellow at the Center for Research on Computation and Society at SEAS and an expert in computer forensics, notes that biology still has a long way to go in its under-standing of genetics, and continuing progress will depend, in part, on the ready availability of a large data set. But individuals may only be willing to contribute to the “genetic commons” if they feel they are receiving something in return—at the very least, protection from harm.

Garfinkel went down the genetic rabbit hole himself, submitting his own sample to a testing site—and, like many others, ended up with more questions than definitive answers.

Privacy is one issue; the validity of the test results is another. A third concern is that the interpreters (whether human or algorithmic) may end up wielding more control than the subjects when it comes to understanding genetic destiny.

In fact, genetic privacy and autonomy are just two threads in a worldwide conversation about technology, society, and the role of (presumed) authorities. Similar issues arise in finance, bioengineering, and cybersecurity. The technology changes continuously, but can policy, ethics, and law keep pace?

The “Triple Academies” meeting I hosted in April represented academic convergence at its finest, something that we at SEAS and Harvard have an opportunity to produce in an unprecedented way. With a few phone calls and emails, we brought together the finest minds on campus in computer science, medicine, genetics, and ethics—along with scholars, leaders, and entrepreneurs from across the country and around the world.

Many questions remain.

Can health privacy legislation like GINA and HIPAA actually keep us safe in a changing world, or does it just help us to feel safe? Are there better options?

As our networked world becomes more saturated with avenues for data sharing and communication, can we trust that our security algorithms will always be strong enough when it counts?

The aim is not to solve these issues once and for all, but to lay the groundwork to contend with what may arise.

While we cannot—and should not—stop technological progress, we have a responsibility as educators and leaders to decide how it will take shape, because it will undoubtedly shape us in return.

Down the Genetic Rabbit Hole

Dean Cherry A. Murray

eliz

a gr

inne

ll

Cherry A. Murray

Dean, Harvard School of Engineering and Applied Sciences John A. and Elizabeth S. Armstrong Professor of Engineering and Applied Sciences Professor of Physics

camurray@seas.harvard.edu

End Note

Kavli visit kicks off new lecture series

In May, we welcomed a visit by Fred Kavli, founder and benefactor of the Kavli Foundation, which funds the Kavli Institute for Bionano Science and Technology at Harvard. As a symbol of gratitude for his recent donation of a faculty chair, Dean Cherry A. Murray presented Kavli with a real hardwood chair.

The first talk in the new Kavli Lecture Series discussed the creation of bio-inspired nanomaterials that mimic the adaptive coloration of cephalopods like squid and cuttlefish.

Harvard to celebrate its 375th anniversary

On October 14 at 7pm in Tercentenary Theatre, Harvard students, faculty, staff, and alumni are cordially invited to gather for a festive evening to include refreshments, colorful parades, and a program with lighting effects and performances by the Harvard/Radcliffe Orchestra, the combined Holden Choirs, and famed cellist Yo-Yo Ma.

Applied math alumna and celebrity pastry chef Joanne Chang ’91 will provide a special birthday cake, and other dessert selections will be available.

SEAS will also mark the anniversary with a special Networks event featuring a Science & Cooking lecture and demonstration, open to all.

Microfluidics lab opens for undergrads

Thanks to a generous gift from Warren Wilkinson ’41, SEAS has opened a new undergraduate teaching lab for microfluidics.

Providing a core facility for students in bioengineering and mechanical engineering to study both fluid dynamics and clinical applications, the lab features state-of-the-art microfluidic pumps, microscopes, ovens, and soft lithography and fabrication equipment.

Support and Engagement

Around Oxford Street

Discover More Online http://seas.harvard.edu/topics

Anonymity, HIPAA, and the U.S. Supreme Court—A closer look at Latanya Sweeney’s research on data re-identification

Q&A with Steven Salzberg, Ph.D. ’89, Professor of Medicine and Biostatistics at Johns Hopkins University and an advocate of open-source genomic research

Q&A with Rachel Greenstadt, Ph.D. ’07, Assistant Professor of Computer Science at Drexel University, whose research involves intelligent, secure systems

“Disappointingly Average”—Author Simson Garfinkel describes his own experience with personal genetic testing

“Triple Academies” video of both expert panel discussions

Additional Resources

caro

line

per

ry

Social Networking

facebook.com/hseas

twitter.com/hseas

HarvardSEAS.tumblr.com

stumbleupon.com, user hseas

29 O

xfor

d St

reet

Ca

mbr

idge

, M

A 02

138

A pu

blic

atio

n of

the

Har

vard

Sch

ool of

Eng

inee

ring

an

d Ap

plie

d Sc

ienc

es C

omm

unic

atio

ns O

ffice

.

Plea

se s

end

feed

back

to:

com

mun

icat

ions

@se

as.h

arva

rd.e

du

or c

all us

at 61

7-49

6-38

15.

Com

mun

icat

ions

Dir

ecto

r M

icha

el P

atrick

Rut

ter

Com

mun

icat

ions

Pro

ject

Man

ager

, Des

igne

r, P

hoto

grap

her

Eliz

a Grinn

ell

Com

mun

icat

ions

Adm

inis

trat

or

Caro

line

Perry

Des

ign

and

Prod

ucti

on

Alp

habe

tica

, In

c.

Copy

edit

or /

Pro

ofre

ader

Ja

mes

Cly

de S

ellm

an, Ph

.D. ’9

3

Feat

ure

Stor

y Au

thor

Si

mso

n L.

Gar

finke

l is

an A

ssoc

iate

Pro

fess

or a

t th

e Nav

al P

ostg

radu

ate

Scho

ol in

Mon

tere

y, C

alifo

rnia

. His

res

earc

h in

tere

sts

incl

ude

com

pute

r fo

rens

ics,

the

em

ergi

ng fie

ld o

f us

abili

ty a

nd s

ecur

ity, p

erso

nal in

for-

mat

ion

man

agem

ent, p

riva

cy, in

form

atio

n po

licy, a

nd ter

rorism

.

Gar

finke

l is

als

o a

jour

nalis

t an

d si

nce

1983

has

written

mor

e th

an a

th

ousa

nd a

rtic

les

on s

cien

ce, te

chno

logy

, an

d te

chno

logy

pol

icy

for

the

popu

lar pr

ess.

He

rece

ived

thr

ee B

ache

lor of

Sci

ence

deg

rees

fro

m M

IT in

1987

, a

Mas

ter’s

of Sc

ienc

e in

Jou

rnal

ism

fro

m C

olum

bia

Uni

vers

ity

in 1

988,

an

d a

Ph.D

. in

Com

pute

r Sc

ienc

e from

MIT

in

2005

. He

com

plet

ed h

is

post

doct

oral

wor

k at

SEA

S, in

asso

ciat

ion

with

the

Ber

kman

Cen

ter.

© C

opyr

ight

by

the

Pres

iden

t an

d Fe

llow

s of

Har

vard

Col

lege

.

ww

w.s

eas.

harv

ard.

edu