MIT Research News' Journal
[Most Recent Entries]
[Calendar View]
Thursday, September 8th, 2016
| Time |
Event |
| 12:00a |
Surrounded by questions Most PhD projects require great commitment, and Nils Wernerfelt's is no exception. Last year, when he suddenly received the opportunity to collaborate with researchers at Facebook, he flew out on a week's notice to San Francisco and has been living out of his suitcase, sleeping on friends’ floors or in Airbnb’s and sublets ever since.
Embedding within a large tech company, working face-to-face with computer scientists, sociologists, and statisticians from around the world has been “really exciting,” Wernerfelt says.
This diversity of backgrounds is particularly appealing to him, as throughout his time at MIT, he has worked on projects ranging from the connection between asthma and vitamin D deficiency during pregnancy, to changing attitudes about the LGBT community. For Wernerfelt, a PhD candidate in economics, there is no shortage of research topics in the world; the difficulty is in deciding which one to pursue next.
Finding questions everywhere
After finishing his bachelor’s degree in mathematics at Harvard University, Wernerfelt knew he wanted to continue on to graduate school, but he wasn’t exactly sure what to study. He applied to a mixture of schools, or, as he puts it, “I threw a bunch of stuff at the wall and saw what stuck,” and eventually decided on economics, a field that appeals to him because it can be applied to so many different subjects.
“I view it as a tool-based discipline,” he says. “It's very flexible in that so many of the things you think about in the world can be understood through an economic lens and then analyzed with the toolkit that you learn.”
For Wernerfelt, the broad applicability of economics has allowed him to continuously mine the world around him for research topics.
“How I approach questions is, you're reading The New York Times, or you're at a bar having a discussion with someone, and you just sort of think of something,” he says. “And then you're like, well, we can test this, let me go get data.”
This is almost exactly how Wernerfelt’s first research project at MIT developed. It started when Richard Zeckhauser, a professor at the Harvard Kennedy School who had worked with Wernerfelt in the past, started chatting with a doctor at a dinner party. The doctor brought up his hypothesis that in utero vitamin D deficiency causes babies to have a higher likelihood of developing asthma. Zeckhauser realized that it might be cheaper and easier to test this hypothesis with an economics approach than with clinical trials. Soon afterward he emailed Wernerfelt, who was immediately intrigued.
“I had never really thought much about asthma before,” he says. “But when we heard about this hypothesis, it was like, wow, here's something that is really impactful that we could actually use economics to say something useful about. About one in 12 Americans has asthma, and it is a hugely costly disease that we don't know much about.”
Wernerfelt dove into the research, focusing on two main datasets. The first was historical data from the Centers for Disease Control and Prevention about when and where people were born, and whether or not they have asthma. The second was data from weather stations around the country that record hours of sunlight every day. (Sunlight accounts for 90 percent of our vitamin D intake).
Wernerfelt realized that combining these datasets would allow him to control for all variables other than sun exposure — for example, he could compare asthma rates of people born in Massachusetts in February of a cloudier year, to asthma rates of people born in Massachusetts in February of a sunnier year. The results indicated that women exposed to more sunlight during their second trimester of pregnancy were significantly less likely to have babies that develop asthma, a finding that is consistent with suggestive evidence from previous medical studies.
A personal quest for answers
It was a much more personal experience, however, that prompted what would become Wernerfelt’s dissertation research. Wernerfelt, who is openly gay, describes himself as “very closeted” during his first few years at Harvard, in part because he didn’t know any gay people. That all changed when he became Facebook friends with a classmate he met in the dining hall.
“He changed his ‘interested in’ field on Facebook to say that he was gay, and he wrote a post about this,” recalls Wernerfelt. “I didn't know any gay people, so this was all of a sudden someone that, because of Facebook, I was able to connect with. Ultimately that first connection gave me the confidence to start down the road to eventually coming out.”
In the years that followed, Wernerfelt noticed that there seemed to be an accelerating wave of support for the LGBT community as more public figures came out and more states legalized same-sex marriage. For example, between the early 1980s and 2013, the number of poll respondents who indicated they personally knew a gay person increased from 24 to 87 percent.
Wernerfelt wanted to study this shift and better understand how attitudes about the LGBT community are changing, which is when he turned back to Facebook.
“I felt like Facebook was the only entity in the world that had data that could speak to this,” he says. “There are a few government surveys that monitor sexual orientation, but their sample sizes are super, super small.”
He points out that Facebook has 1.7 billion monthly active users, and after the Supreme Court ruling on marriage equality, some 26 million users overlaid a rainbow flag on their profile pictures in a 48-hour period. More broadly, Wernerfelt concluded that there is evidence that the connectivity of Facebook is making a difference for the LGBT community.
“For a long time in this country, because there weren’t a lot of openly LGBT people, very few people knew any, and so it was easier for stigma to persist,” he says. “What you see is that support and people coming out are going up in tandem. I think that the personal contact has played a huge role, and that’s something Facebook has really enabled.”
The path forward
Wernerfelt was so excited about the research questions he could tackle at Facebook that he has decided to stay at the company in its Core Data Science group.
Wernerfelt, who will finish his PhD this September, greatly credits his advisors Michael Whinston, David Autor, Daron Acemoglu, and Richard Zeckhauser, and his classmates with helping him grow intellectually and personally during his time at MIT. Ultimately, he would like to return to academia, but for the foreseeable future he is excited to pursue whatever questions and opportunities he finds at Facebook.
While this indirect path engenders some uncertainty, he says his current mindset is best embodied in a quotation by poet Ranier Maria Rilke that ends, “Do not now seek the answers, which cannot be given you because you would not be able to live them. And the point is, to live everything. Live the questions now. Perhaps you will then gradually, without noticing it, live along some distant day into the answer.”
“Right now what I want to do is live the questions,” Wernerfelt says. “And someday I'll look back and I'll have a career that will be one I've lived myself into, no doubt centered around studies of economics and human behavior.” | | 12:00a |
MIT’s REXIS is bound for asteroid Bennu An SUV-sized spacecraft, loaded with instruments and an extendable robotic arm, will soon be barreling toward a space rock, on a round-trip journey that promises to return an unprecedented souvenir: extraterrestrial soil, taken directly from an asteroid, that could hold clues to the very early universe.
This evening, at about 7:05 p.m. EDT, NASA’s first-ever asteroid sample return mission, OSIRIS-REx, (Origins, Spectral Interpretation, Resource Identification, Security-Regolith Explorer), will launch from Cape Canaveral, Florida. The spacecraft embarks on a two-year, 120-million-mile haul, at the end of which it will reach the asteroid 101955 Bennu, a near-Earth asteroid that is thought to be a cosmological “time-capsule,” composed of remnants of the early universe.
The spacecraft will spend some time remotely analyzing the asteroid’s surface using a suite of instruments, including REXIS, (Regolith X-ray Imaging Spectrometer), an instrument designed and built by more than 50 MIT students. The spectrometer will analyze the interaction of the sun’s X-rays with the soil, or regolith, to identify chemical elements on Bennu’s surface. REXIS may also help determine the best locations for the robotic arm to reach out and grab a sample. The mission is scheduled to return an asteroid sample to Earth in 2023.
Prior to liftoff, MIT News spoke with Richard Binzel, professor of earth, atmospheric, and planetary sciences at MIT, and principal investigator for REXIS, about MIT’s role in launching the mission.
Q: While OSIRIS-REx has just begun its journey to Bennu, it’s almost as if another journey has ended — namely, the efforts by hundreds of scientists and students to get this mission off the ground. Can you take us through some of the highlights, scientific and personal, from the final weeks of preparing for the launch?
A: It may be hard to believe, but for our MIT team and for our partners at Harvard, the final countdown began five years ago. Outer space knows no weekends, no holidays, only the relentless march toward the date when the Earth and our target asteroid are in perfect alignment for launch. Every decision we have made toward designing, building, and testing the spacecraft and its instruments had to be evaluated against this immutable schedule. So it has been a five-year marathon at full sprint knowing that you face Heartbreak Hill every day. You are always one slip, one faulty part, one failed test away from being knocked out of the race against the clockwork of the solar system.
So, finally getting to the launch pad feels surreal. You see the finish line in front of you but suddenly everything seems to decelerate into slow motion. Your final spacecraft checks are complete, and the focus shifts from the spacecraft itself to the launch team whose job is to get you off the pad. All you can do is wait and for once, you actually want the clock to tick down to zero. But as you wait, you are still constantly going back through your head and wondering, “What have we overlooked? What have we failed to anticipate?” There is always something unexpected, but we hope we have built in enough extra margin of capability to deal with it.
Launch is a highlight not only for being a spectacular sight, but that final countdown moment brings all of the team together to a single place and time. We have all been working hard at our own facilities and universities, but now we are all together to share a great moment of success. Our REXIS instrument on board OSIRIS-REx has involved more than 50 students, many of whom don’t know one another because they graduated and moved on before the next generation arrived. Seeing those past and present students come together and connect through their shared experience makes me very proud of our team. It is amazing how much they have accomplished in what seems like a very short time.
Q: The main goal of the mission is to pick a sample from the asteroid and return it to Earth. To accomplish this, what are the main steps the spacecraft will have to take once it arrives at the asteroid, and what role will MIT’s REXIS instrument play in this endeavor?
A: From the launch pad we have a two-year journey around the sun to catch up to Bennu and fire our thrusters to have the OSIRIS-REx spacecraft go into orbit. Our first orbits are rather distant as we get the “feel” for Bennu’s gravity field before going lower. We have a suite of cameras that will begin a detailed reconnaissance of the asteroid surface to create a list of promising sample sites. Safety is the first priority for any site. Basically we want large, smooth areas that aren’t filled with craggy boulders.
When OSIRIS-REx descends into its second, lower orbit, that’s when the REXIS instrument swings into action. Literally! We have to execute a command that enables our spring-loaded cover to pop open. That cover protects our detectors from space radiation during our two-year flight to Bennu. Once we are open, our first step is to point out to space at a well-known astronomical source so that we can calibrate how well our instrument is functioning. Getting that cover open and those first confirming measurements will be heart-stopping moments for us.
REXIS has “X-ray eyes” that will then begin scanning Bennu’s surface to map out regions that may be glowing (fluorescing) in response to the sun’s X-ray light streaming onto its surface. By measuring the exact energies that the surface emits, we can find regions that might be particularly rich in iron, silicon, sulfur, and magnesium. We will have our map of atomic elements to compare with other teams’ maps of minerals, and we will collectively weigh the science pros and cons of each site. Our emphasis will be on material that looks like it has been well-preserved since the beginning of our solar system. Once selected, OSIRIS-REx will descend like a giant mosquito, with its long arm touching the surface for just a few seconds to snatch our sample. With that sample secured into a re-entry capsule, it will be a long journey back to Earth. The journey ends dramatically in 2023 when the OSIRIS-REx spacecraft drops off that capsule back into the atmosphere. Seeing the parachutes deploy and that precious cargo landing safely will be a great triumph.
Q: What will MIT’s role be once the spacecraft returns the sample to Earth?
A: At MIT we have some of the most sophisticated sample-analysis instruments on Earth, where even a fraction of a gram of new material is a treasure trove of new information. That’s why we are working so hard to get a pristine sample from space directly into our laboratories. Our first priority is to look at the chemistry of the sample, because we think Bennu is like a Rosetta stone for the original material from which the Earth formed and life evolved. It is a way of looking back in time at our very own human beginnings. From there, we will want to decode the history of the solar system, as Bennu itself is a surviving piece of the building blocks of planets.
For the REXIS team, we are very interested to see if the elemental chemistry we “predicted” with our X-ray measurements at Bennu proves to be correct. It’s not just because we care about getting an A+. It’s because demonstrating that a REXIS-like instrument is a reliable extension of human capability creates an enabling technology we can apply to countless new destinations. There are thousands of asteroids in orbits accessible from Earth that we care about for science, for future resources in space, and perhaps even to evaluate some future asteroid impact hazard. If a shoebox-sized instrument like REXIS can prove itself to be a trusty explorer, we are opening a gateway to a long-term future of ongoing exploration. That is the legacy that I hope REXIS will leave and that every student involved can be very proud of. | | 12:00p |
How the brain builds panoramic memory When asked to visualize your childhood home, you can probably picture not only the house you lived in, but also the buildings next door and across the street. MIT neuroscientists have now identified two brain regions that are involved in creating these panoramic memories.
These brain regions help us to merge fleeting views of our surroundings into a seamless, 360-degree panorama, the researchers say.
“Our understanding of our environment is largely shaped by our memory for what’s currently out of sight,” says Caroline Robertson, a postdoc at MIT’s McGovern Institute for Brain Research and a junior fellow of the Harvard Society of Fellows. “What we were looking for are hubs in the brain where your memories for the panoramic environment are integrated with your current field of view.”
Robertson is the lead author of the study, which appears in the Sept. 8 issue of the journal Current Biology. Nancy Kanwisher, the Walter A. Rosenblith Professor of Brain and Cognitive Sciences and a member of the McGovern Institute, is the paper’s lead author.
Building memories
As we look at a scene, visual information flows from our retinas into the brain, which has regions that are responsible for processing different elements of what we see, such as faces or objects. The MIT team suspected that areas involved in processing scenes — the occipital place area (OPA), the retrosplenial complex (RSC), and parahippocampal place area (PPA) — might also be involved in generating panoramic memories of a place such as a street corner.
If this were true, when you saw two images of houses that you knew were across the street from each other, they would evoke similar patterns of activity in these specialized brain regions. Two houses from different streets would not induce similar patterns.
“Our hypothesis was that as we begin to build memory of the environment around us, there would be certain regions of the brain where the representation of a single image would start to overlap with representations of other views from the same scene,” Robertson says.
The researchers explored this hypothesis using immersive virtual reality headsets, which allowed them to show people many different panoramic scenes. In this study, the researchers showed participants images from 40 street corners in Boston’s Beacon Hill neighborhood. The images were presented in two ways: Half the time, participants saw a 100-degree stretch of a 360-degree scene, but the other half of the time, they saw two noncontinuous stretches of a 360-degree scene.
After showing participants these panoramic environments, the researchers then showed them 40 pairs of images and asked if they came from the same street corner. Participants were much better able to determine if pairs came from the same corner if they had seen the two scenes linked in the 100-degree image than if they had seen them unlinked.
Brain scans revealed that when participants saw two images that they knew were linked, the response patterns in the RSC and OPA regions were similar. However, this was not the case for image pairs that the participants had not seen as linked. This suggests that the RSC and OPA, but not the PPA, are involved in building panoramic memories of our surroundings, the researchers say.
Priming the brain
In another experiment, the researchers tested whether one image could “prime” the brain to recall an image from the same panoramic scene. To do this, they showed participants a scene and asked them whether it had been on their left or right when they first saw it. Before that, they showed them either another image from the same street corner or an unrelated image. Participants performed much better when primed with the related image.
“After you have seen a series of views of a panoramic environment, you have explicitly linked them in memory to a known place,” Robertson says. “They also evoke overlapping visual representations in certain regions of the brain, which is implicitly guiding your upcoming perceptual experience.”
The research was funded by the National Science Foundation Science and Technology Center for Brains, Minds, and Machines; and the Harvard Milton Fund. | | 12:00p |
Linking RNA structure and function Several years ago, biologists discovered a new type of genetic material known as long noncoding RNA. This RNA does not code for proteins and is copied from sections of the genome once believed to be “junk DNA.”
Since then, scientists have found evidence that long noncoding RNA, or lncRNA, plays roles in many cellular processes, including guiding cell fate during embryonic development. However, it has been unknown exactly how lncRNA exerts this influence.
Inspired by historical work showing that structure plays a role in the function of other classes of RNA such as transfer RNA, MIT biologists have now deciphered the structure of one type of lncRNA and used that information to figure out how it interacts with a cellular protein to control the development of heart muscle cells. This is one of first studies to link the structure of lncRNAs to their function.
“Emerging data points to fundamental roles for many of these molecules in development and disease, so we believe that determining the structure of lncRNAs is critical for understanding how they function,” says Laurie Boyer, the Irwin and Helen Sizer Career Development Associate Professor of Biology and Biological Engineering at MIT and the senior author of the study, which appears in the journal Molecular Cell on Sept. 8.
Learning more about how lncRNAs control cell differentiation could offer a new approach to developing drugs for patients whose hearts have been damaged by cardiovascular disease, aging, or cancer.
The paper’s lead author is MIT postdoc Zhihong Xue. Other MIT authors are undergraduate Boryana Doyle and Sarnoff Fellow Arune Gulati. Scott Hennelly, Irina Novikova, and Karissa Sanbonmatsu of Los Alamos National Laboratory are also authors of the paper.
Probing the heart
Boyer’s lab previously identified a mouse lncRNA known as Braveheart, which is found at higher levels in the heart compared to other tissues. In 2013, Boyer showed that this RNA molecule is necessary for normal development of heart muscle cells.
In the new study, the researchers decided to investigate which regions of the 600-nucleotide RNA molecule are crucial to its function. “We knew Braveheart was critical for heart muscle cell development, but we didn’t know the detailed molecular mechanism of how this lncRNA functioned, so we hypothesized that determining its structure could reveal new clues,” Xue says.
To determine Braveheart’s structure, the researchers used a technique called chemical probing, in which they treated the RNA molecule with a chemical reagent that modifies exposed RNA nucleotides. By analyzing which nucleotides bind to this reagent, the researchers can identify single-stranded regions, double-stranded helices, loops, and other structures.
This analysis revealed that Braveheart has several distinct structural regions, or motifs. The researchers then tested which of these motifs were most important to the molecule’s function. To their surprise, they found that removing 11 nucleotides, composing a loop that represents just 2 percent of the entire molecule, halted normal heart cell development.
The researchers then searched for proteins that the Braveheart loop might interact with to control heart cell development. In a screen of about 10,000 proteins, they discovered that a transcription factor protein called cellular nucleic acid binding protein (CNBP) binds strongly to this region. Previous studies have shown that mutations in CNBP can lead to heart defects in mice and humans.
Further studies revealed that CNBP acts as a potential roadblock for cardiac development, and that Braveheart releases this repressor, allowing cells to become heart muscle.
“This is one of the first studies to relate lncRNA structure to function,” says John Rinn, a professor of stem cell and regenerative biology at Harvard University, who was not involved in the research.
“It is critical that we move toward understanding the specific functional domains and their structural elements if we are going to get lncRNAs up to speed with proteins, where we already know how certain parts play certain roles. In fact, you can predict what a protein does nowadays because of the wealth of structure-to-function relationships known for proteins,” Rinn says.
Building a fingerprint
Scientists have not yet identified a human counterpart to the mouse Braveheart lncRNA, in part because human and mouse lncRNA sequences are poorly conserved, even though protein-coding genes of the two species are usually very similar. However, now that the researchers know the structure of the mouse Braveheart lncRNA, they plan to analyze human lncRNA molecules to identify similar structures, which would suggest that they have similar functions.
“We’re taking this motif and we’re using it to build a fingerprint so we can potentially find motifs that resemble that lncRNA across species,” Boyer says. “We also hope to extend this work to identify the modes of action of a catalog of motifs so that we can better predict lncRNAs with important functions.”
The researchers also plan to apply what they have learned about lncRNA toward engineering new therapeutics. “We fully expect that unraveling lncRNA structure-to-function relationships will open up exciting new therapeutic modalities in the near future,” Boyer says. | | 2:00p |
Genomes, good news, and you Personal genomics lets people evaluate their risk levels for many diseases. But how do they respond to that data? A unique new study co-authored by investigators from MIT and Brigham and Women’s Hospital has some answers: People react more strongly to good medical news than to bad medical news, and they respond more to one surprising result than to a broad array of findings.
The study, one of the first of its kind, uses before-and-after evaluations of risk by consumers regarding eight serious diseases, including cancer, heart disease, Alzheimer’s disease, and Parkinson’s disease. The research examines subjective reactions and looks at the extent to which people change their self-assessments of their risk of developing an illness, based on new genetic information.
“If people get good news, they significantly decrease their risk perception,” says Joshua Krieger, a doctoral candidate at the MIT Sloan School of Management, and lead author of a new paper detailing the study. “You see a much bigger change for good news than bad news.”
In fact, people adjust their risk perceptions about twice as much for positive news as for negative news.
“There is an optimism bias, in that people want to believe good news,” says Krieger, drawing on related research in cognitive psychology.
“This research expands our understanding of how the consumers of personal genomics may react to what they perceive as elevated or diminished risk, and reminds us that perceptions are not always logical,” says Robert C. Green, a medical geneticist at Brigham and Women’s Hospital and Harvard Medical School.
The study is part of the long-term Impact of Personal Genomes (PGen) project, which examines many aspects of the field. Its principal investigators are Green, and J. Scott Roberts, an associate professor at the University of Michigan School of Public Health.
The paper, “The impact of personal genomics on risk perceptions and medical decision-making,” is published today in Nature Biotechnology.
The authors are Krieger; Fiona Murray, who is associate dean for innovation, co-director of the MIT Innovation Initiative, and the William Porter (1967) Professor of Entrepreneurship at MIT Sloan; Roberts; and Green, a physician and genetics researcher who is also a faculty member at the Broad Institute.
Policy shift
The study looked at the responses of 617 consumers who received genomic data from the personal genomics company 23andMe. The respondents reported their risk perception for eight medical conditions before receiving the results, and then again six months after receiving the results. The research is the first one to evaluate consumers’ responses in a real-world setting, based on data from a commercial firm.
The diseases in the study are lung cancer, Parkinson’s disease, colorectal cancer, Alzheimer’s disease, Type II diabetes, breast cancer, prostate cancer, and coronary heart disease.
In the time since the study was conducted, the U.S. Food and Drug Administration (FDA) in 2013 required 23andMe to stop offering direct-to-consumer genetics tests for disease risks. (Such firms also provide consumers with ancestry information.) However, in 2015 the FDA agreed to let 23andMe offer a smaller set of medical results to consumers.
The study also examined the extent to which changes in risk perception, based on the new genomic data, might spur people into making follow-up visits to medical providers, and under what circumstances.
“People aren’t responding to the trends across all their disease [data], but they are willing to take follow-up actions when they get one or two results that are really surprising,” Krieger adds.
Krieger also cites the work of Tali Sharot, a cognitive neuroscientist at University College London and a recent visiting scientist at MIT, whose research on “optimism bias” helped provide groundwork for the researchers’ interpretation of the results.
Following up, but not flocking
When consumers evaluated their perceptions on a five-point scale, a one-point change in terms of increased risk across all eight disease categories made them 5 percent more likely to make a follow-up medical appointment, and 4 percent more likely to have follow-up exams or procedures.
That suggests people might not necessarily swamp doctors with requests for examinations after receiving genetic information on disease risks, one of the concerns surrounding the direct-to-consumer genetic testing business.
“It doesn’t seem like people are flocking to the doctor with [insignificant] results and trying to take action with every little change in risk perception,” Krieger says.
Bad news about genetic risks also led to a larger change in risk perception for Alzheimer’s disease than for the other diseases in the study. Such variations in patient reactions, Krieger suggests, could make policy development especially complex in personal genomics.
There may not be “a one-size-fits-all approach to regulating [information about] genetic diseases,” Krieger observes. “Diseases vary in their severity, their mortality, their ability for us to treat those diseases, and it seems like those characteristics matter for patients as well.”
The study was supported by the National Institutes of Health. |
|