Data Management & Statistics

The rigors of an MIT education inevitably lead to academic setbacks. Now, students are also navigating the social isolation of remote learning compounded with the stress of a global pandemic and tense political climate.
On Jan. 21, MIT students gathered virtually for a screening of the “Flipping Failure COVID Stories,” followed by a moderated panel discussion with their filmmaking peers. In the independently-produced short films, the students document the internal struggles that they are dealing with during the pandemic.  
The first-time filmmakers expressed a shared — but often unspoken — experience of feeling inadequate or “stuck” when measuring themselves against their ambitions or expectations. However, through the creative process of transforming their self-observations into art, they furthered their journeys toward self-acceptance and self-compassion, while normalizing the experience of struggling for their peers. 
Lulu Tian, a junior in MIT’s Department of Brain and Cognitive Sciences (BCS), captures her transition from living on campus to at home with her family in her film “Re-remembering.” She sets her experience of reconnecting with herself against a backdrop of uncertainty surrounding the pandemic. In a familiar scene, a family member cooks while Anderson Cooper announces that the crisis “might get worse before it gets better” on a nearby screen.
“This is what it feels like to start again,” narrates Tian, as she chips off dried paint from a palette to revisit painting as a creative outlet. Painting is also used as a visual metaphor for creative ambition in another short, “Loop(s),” by Maia Campbell, a fellow junior in BCS. She narrates over a time-lapse of her artwork, “Imagine if I were to really work on something for a long time in a meaningful way. What would that thing be that I create?” 
Both students express a value embedded in MIT’s culture, as emblematic as its beaver mascot: the drive to build and create something impactful. Still, this aspiration can become a stressful expectation that students impose on themselves during a time when they, and the world, are living through a pandemic. 
Whether students frame their academic or personal outcomes as “failures” hinges on their perspective. Campbell considers the tension between her feelings of guilt over the paralysis she feels during the pandemic and the self-compassion she knows she should be practicing. “I feel like it’s advice that I give, but I don’t follow,” she says. 
Andy Haupt, a PhD student in the MIT Institute for Data, Systems, and Society, similarly gave voice to the social isolation and indecision experienced by many students indefinitely relocated off-campus in the film, “Feeling Stuck… A Bit.” “I find art a great language to express things and to understand things which are hard to understand,” he says. “I wanted to make a video about my struggles and my experiences through the time of Covid.”
The short films provide insight into how students can adapt to a pervasive disruption. In the panel discussion following the screening, Haupt reflected on how he coped with his loss of agency in one area, where he lived and worked, by focusing instead on the importance of who he worked with and other positive aspects of his graduate experience. Others reflected that the setback prompted them to reconnect with their values and find more space for themselves. 
Stories of resilience
“The COVID Diaries” are the most recent contribution to the Flipping Failure initiative, a collection of stories told by undergraduate and graduate students about their journeys toward academic resilience and self-acceptance at MIT. The initiative is facilitated by MIT’s Teaching + Learning Lab (TLL), which is interested in humanizing experiences of academic struggle and sharing student’s coping strategies.
Viewers may find that they have experienced similar challenges: questioning their abilities after under-performing on an exam, managing work overload, or struggling to find their academic or career path. The site aims to highlight that, while one might feel alone in their struggle, MIT peers may have experienced related difficulties, too. Students also share positive coping strategies, such as how to adopt a growth mindset, reframe negative thoughts, improve studying strategies, build a support network, and elicit guidance from mentors.
“The student contributors to Flipping Failure are an amazing group. As we constantly see with MIT students, they have such a strong desire to help their peers. To put their stories out there — to show that kind of vulnerability — is so brave,” says Dipa Shah, senior associate director in TLL. 
Through their vulnerability, student contributors hope to provide the reassurance needed for others to persevere through similar challenges. “It made me think deeply about how important it is to be open about our struggles and failures so that those going through those same periods of self-doubt do not feel like it is unique to them,” says contributor Jennifer Cherone, a PhD student in the Department of Biology.
Shedding the stigma
Research indicates that students’ beliefs about the nature of intelligence and ability significantly shape their response to academic challenges. Students who adopt a growth mindset allocate more effort, experiment with new approaches, and actively seek feedback. 
“Struggle comes as a result of learning and growing, and is how we build resilience,” says Lourdes Aleman, associate director in TLL. “Many MIT students encounter their first real academic struggles at MIT, which often results in feelings of inadequacy. Often, students incorrectly assume they are the only ones struggling and experience shame and stress as a consequence.”
To foster a more growth-minded culture for students, Flipping Failure aims to “flip” the conversation that labels academic setbacks as failures and re-frame them as an expected part of the human experience. Such a paradigm shift can improve students’ academic experiences and help alleviate related stress.
The process of contributing to Flipping Failure is itself transformative for students. In creating a safe space for exchanging stories of the academic challenges they have faced and their learning experiences, student contributors quickly become aware they are far from alone. 
“The Flipping Failure project has helped me reflect on my experiences, vocalize and internalize the lessons learned from these experiences, and understand how my different strategies and techniques string together to help me become more accepting of myself,” says contributor Kanika Gakhar SM ’20, who recently graduated with a master’s degree from the Department of Aeronautics and Astronautics. “By sharing my experience and strategies, I can forgive myself a little more and find comfort in knowing that not only am I learning from my failures, but so are others.”
One MIT
Members of the MIT community have a history of supporting each other and contributing to a positive culture for growth. The Flipping Failure project was inspired by past resilience-building projects within our community, such as Portraits of Resilience, as well as similar initiatives at peer institutions.
Flipping Failure also complements more recent efforts on campus to help students manage and overcome setbacks, such as FAIL! and the Undergraduate Association’s MIT I Messed Up events. More

The specter of “fake news” looms over many facets of modern society. Waves of online misinformation have rocked societal events from the Covid-19 pandemic to U.S. elections. But it doesn’t have to be that way, according to Devavrat Shah, a professor in the Department of Electrical Engineering and Computer Science and the Institute for Data, Systems and Society. Shah researches the recommendation algorithms that generate social media newsfeeds. He has proposed a new approach that could limit the spread of misinformation by emphasizing content generated by a user’s own contacts, rather than whatever happens to be trending globally. As Congress and a new presidential administration mull whether and how to regulate social media, Shah shared his thoughts with MIT News.
Q: How does misinformation spread online, and do social media algorithms accelerate that spread?
A: Misinformation spreads when a lie is repeated. This goes back thousands of years. I was reminded last night as I was reading bedtime stories to my 6-year-old, from the Panchatantra fables:
A brahmin once performed sacred ceremonies for a rich merchant and got a goat in return. He was on his way back carrying the goat on his shoulders when three crooks saw him and decided to trick him into giving the goat to them. One after the other, the three crooks crossed the brahmin’s path and asked him the same question – “O Brahmin, why do you carry a dog on your back?”
The foolish Brahmin thought that he must indeed be carrying a dog if three people have told him so. Without even bothering to look at the animal, he let the goat go.
In some sense, that’s the standard form of radicalization: You just keep hearing something, without question and without alternate viewpoints. Then misinformation becomes the information. That is the primary way information spreads in an incorrect manner. And that’s the problem with the recommendation algorithms, such as those likely to be used by Facebook and Twitter. They often prioritize content that’s gotten a lot of clicks and likes — whether or not it’s true — and mixes it with content from sources that you trust. These algorithms are fundamentally designed to concentrate their attention onto a few viral posts rather than diversify things. So, they are unfortunately facilitating the process of misinformation.
Q: Can this be fixed with better algorithms? Or are more human content moderators necessary?
A: This is doable through algorithms. The problem with human content moderation is that a human or tech company is coming in and dictating what’s right and what’s wrong. And that’s a very reductionist approach. I think Facebook and Twitter can solve this problem without being reductionist or having a heavy-handed approach in deciding what’s right or wrong. Instead, they can avoid this polarization and simply let the networks operate the way the world operates naturally offline, though peer interactions. Online social networks have twisted the flow of information and put the ability to do so in the hands of a few. So, let’s go back to normalcy.
There’s a simple tweak that could make an impact: A measured amount of diversity should be included in the newsfeeds by all these algorithms. Why? Well, think of a time before social media, when we may chat with people in an office or learn news through friends. Although we are still exposed to misinformation, we know who told us that information, and we tend to share it only if we trust that person. So, unless that misinformation comes from many trusted sources, it is rarely widely shared.
There are two key differences online. First, the content that platforms insert is mixed in with content from sources that we trust, making it more likely for us to take that information at face value. Second, misinformation can be easily shared online so that we see it many times and become convinced it is true. Diversity helps to dilute misinformation by exposing us to alternate points of view without abusing our trust. 
Q: How would this work with social media?
A: To do this, the platforms could randomly subsample posts in a way that looks like reality. It’s important that a platform is allowed to algorithmically filter newsfeeds — otherwise there will be too much content to consume. But rather than rely on recommended or promoted content, a feed could pull most of its content, totally at random, from all of my connections on the network. So, content polarization through repeated recommendation wouldn’t happen. And all of this can — and should — be regulated.
One way to make progress toward more natural behavior is by filtering according to a social contract between users and platforms, an idea legal scholars are already working on. As we discussed, the newsfeed of users impacts their behaviors, such as their voting or shopping preferences. In a recent work, we showed that we can use methods from statistics and machine learning to verify whether or not the filtered newsfeed respects the social contract in terms of how it affects user behaviors. As we argue in this work, it turns out that such contracting may not impact the “bottom line” revenue of the platform itself. That is, the platform does not necessarily need to choose between honoring the social contract and generating revenue.
In a sense, other utilities like the telephone service providers are already obeying this kind of contractual arrangement with the “no spam call list” and by respecting whether your phone number is listed publicly or not. By distributing information, social media is also providing a public utility in a sense, and should be regulated as such. More

Researchers have developed a data transfer system that can transmit information 10 times faster than a USB. The new link pairs high-frequency silicon chips with a polymer cable as thin a strand of hair. The system may one day boost energy efficiency in data centers and lighten the loads of electronics-rich spacecraft.
The research was presented at this month’s IEEE International Solid-State Circuits Conference. The lead author is Jack Holloway ’03, MNG ’04, who completed his PhD in MIT’s Department of Electrical Engineering and Computer Science (EECS) last fall and currently works for Raytheon. Co-authors include Ruonan Han, associate professor and Holloway’s PhD adviser in EECS, and Georgios Dogiamis, a senior researcher at Intel.
The need for snappy data exchange is clear, especially in an era of remote work. “There’s an explosion in the amount of information being shared between computer chips — cloud computing, the internet, big data. And a lot of this happens over conventional copper wire,” says Holloway. But copper wires, like those found in USB or HDMI cables, are power-hungry — especially when dealing with heavy data loads. “There’s a fundamental tradeoff between the amount of energy burned and the rate of information exchanged.” Despite a growing demand for fast data transmission (beyond 100 gigabits per second) through conduits longer than a meter, Holloway says the typical solution has been “increasingly bulky and costly” copper cables.
One alternative to copper wire is fiber-optic cable, though that has its own problems. Whereas copper wires use electrical signaling, fiber-optics use photons. That allows fiber-optics to transmit data quickly and with little energy dissipation. But silicon computer chips generally don’t play well with photons, making interconnections between fiber-optic cables and computers a challenge. “There’s currently no way to efficiently generate, amplify, or detect photons in silicon,” says Holloway. “There are all kinds of expensive and complex integration schemes, but from an economics perspective, it’s not a great solution.” So, the researchers developed their own.
The team’s new link draws on benefits of both copper and fiber optic conduits, while ditching their drawbacks. “It’s a great example of a complementary solution,” says Dogiamis. Their conduit is made of plastic polymer, so it’s lighter and potentially cheaper to manufacture than traditional copper cables. But when the polymer link is operated with sub-terahertz electromagnetic signals, it’s far more energy-efficient than copper in transmitting a high data load. The new link’s efficiency rivals that of fiber-optic, but has a key advantage: “It’s compatible directly with silicon chips, without any special manufacturing,” says Holloway.
The team engineered such low-cost chips to pair with the polymer conduit. Typically, silicon chips struggle to operate at sub-terahertz frequencies. Yet the team’s new chips generate those high-frequency signals with enough power to transmit data directly into the conduit. That clean connection from the silicon chips to the conduit means the overall system can be manufactured with standard, cost-effective methods, the researchers say.
The new link also beats out copper and fiber optic in terms of size. “The cross-sectional area of our cable is 0.4 millimeters by a quarter millimeter,” says Han. “So, it’s super tiny, like a strand of hair.” Despite its slim size, it can carry a hefty load of data, since it sends signals over three different parallel channels, separated by frequency. The link’s total bandwidth is 105 gigabits per second, nearly an order of magnitude faster than a copper-based USB cable. Dogiamis says the cable could “address the bandwidth challenges as we see this megatrend toward more and more data.”
In future work, Han hopes to make the polymer conduits even faster by bundling them together. “Then the data rate will be off the charts,” he says. “It could be one terabit per second, still at low cost.”
The researchers suggest “data-dense” applications, like server farms, could be early adopters of the new links, since they could dramatically cut data centers’ high energy demands. The link could also be a key solution for the aerospace and automotive industries, which place a premium on small, light devices. And one day, the link could replace the consumer electronic cables in homes and offices, thanks to the link’s simplicity and speed. “It’s far less costly than [copper or fiber optic] approaches, with significantly wider bandwidth and lower loss than conventional copper solutions,” says Holloway. “So, high fives all round.”
This research was funded, in part, by Intel, Raytheon, the Naval Research Laboratory, and the Office of Naval Research. More

“A seed grant for a risky idea that is mission-driven goes a long way.” 
These are the words of Fadel Adib, an associate professor of media arts and sciences and of electrical engineering and computer science and a 2019 recipient of a two-year seed grant from the Abdul Latif Jameel Water and Food Systems Lab (J-WAFS) at MIT. His work is in wireless sensing, where his research group has largely focused on developing fundamental technology. It is technology with a mission, however one that — until the J-WAFS seed grant — had largely focused on supporting human health and the environment, but not yet food. “I started with an early project applied to food, but the results were not enough to publish. When I saw the J-WAFS seed grant request for proposals I realized that this was a great way for [my research group] to expand our efforts in the food sector.” The resulting research project, a wireless sensor that uses RFID technology to measure the safety and nutritional quality of food and beverage products, has since inspired him and his research group to delve deeper into food-sector research, including exploring potential applications for sustainable aquaculture.
Adib’s story is one of many that J-WAFS principal investigators — especially junior faculty — have shared about the impact of the seed grant program, illustrating how influential this grant can be. Funding is an essential research driver, with the availability of resources often defining the subject area and scope of individual projects, as well as entire careers. Seed grants in particular can be transformational, especially for junior faculty such as Adib, for the opportunity they provide for exploration.
J-WAFS catalyzes research across all disciplines and programs at MIT in order to find solutions to urgent global water and food systems challenges. For MIT faculty coming from technology-dominant disciplines, this emphasis on impact can be invigorating. For Adib, “it allowed [my lab] to do more interdisciplinary work … starting with the problem first, rather than the technique.”
Mathias Kolle, Rockwell Career Development Professor in the Department of Mechanical Engineering and a J-WAFS grantee, agrees. Kolle received a J-WAFS seed grant in 2017 to develop novel, light-diffusing fibers to increase the energy efficiency of industrial algae production in order to improve its viability as an affordable, environmentally sustainable solution for food and fuel. He comments that the grant proposal and review process itself helped him connect the dots between the technology milestones he sought to pursue and the social impact potential of the project. He credits the prompts sent by the expert reviewers convened by J-WAFS in the final stages of the grant process for “helping me create a pretty convincing picture of why work on algae is important.” Joseph Sandt SM ’15, PhD ’20 collaborated with Kolle throughout his PhD program in mechanical engineering, making the project the focus of his thesis. Kolle comments, “Joseph was very fired-up when the J-WAFS project came up.” The research allowed him to build on his existing interest in sustainability while working on an engineering project that still involved a lot of tinkering.
A J-WAFS seed grant inspired yet another junior faculty member to pursue water and food research for the first time: Julia Ortony, the Finmeccanica Assistant Professor in Materials Science and Engineering.   The 2018 grant she received was her first major grant as a new junior faculty member in the Department of Materials Science and Engineering — one that allowed her to work on applied instead of fundamental research for the very first time. “[The J-WAFS seed project] was the first time I really thought about the end product,” she says. Through it, Ortony and her lab develop molecule-based nanofiber hydrogels that are able to bind arsenic and other heavy metals in order to clean drinking water. Ortony recalls, “at the time we received the grant, we were a very new group. It was hard for us to get a big federal grant without preliminary data.”
The J-WAFS grant served as an important catalyst. Data from the J-WAFS project drove another successful grant, the Professor Amar G. Bose Research Grant, which enabled the continuation of the J-WAFS research in a different state of matter. “We wouldn’t have been able to explore solid state nanomaterials without the knowledge we gained from the J-WAFS project,” Ortony comments. Since then, she has received additional follow-on funding in the form of a CAREER award from the National Science Foundation, which will enable her research team to develop their understanding of the fundamentals of the nanofiber materials in order to learn how to tune it to even more effectively pull heavy metal contaminants from water. “The J-WAFS seed grant has allowed our group to make a right turn and think about the goal of our research from an applications perspective,” comments Ortony. “We are now doing this in other domains too, outside of water purification.”
The mission-oriented focus of the J-WAFS seed grant attracted another junior faculty member: Joann de Zegher, the Maurice F. Strong Career Development Professor at the MIT Sloan School of Management. Joann joined MIT in the fall of 2018 after completing a PhD and postdoc at Stanford University. While there, she had been working on the sustainability of global supply chains, focusing on contract design that more effectively aligns incentives with global sourcing and sustainability. Unlike Kolle, Adib, and Ortony, de Zegher had already begun working in the food sector, having pivoted toward understanding supply chain management to support the sustainability of informal food systems. Her 2019 J-WAFS seed grant is supporting the development of mobile supply chain platforms to support sustainable palm oil production by smallholder farmers in Indonesia.
Fieldwork is essential to de Zegher’s research, yet “fieldwork is expensive,” she says, and notes that “When it comes to the study of informal supply chains, like smallholder farmers in Indonesia, it’s hard to find opportunities to fund things like travel and student research assistants.” This is where her 2019 J-WAFS seed grant has proved influential. It “provides an important complement to the funding from foundations that supports field operations.” The J-WAFS funding for travel, fieldwork, and a full-time student supporting data collection and analysis has enabled “that extra-mile data analysis that could have been missing” had de Zegher not received the grant.
The solutions-oriented approach that the seed grant program takes welcomes cross-disciplinary, collaborative approaches to problem-solving. J-WAFS has funded many interdepartmental research collaborations in which junior faculty have been involved. One such collaboration is between David Des Marais, Gale Assistant Professor of Civil and Environmental Engineering, and Caroline Uhler, the Henry L. and Grace Doherty Associate Professor of Electrical Engineering and Computer Science. They are working on a J-WAFS-backed project to find the genetic foundations of plant tolerance to the stresses of heat and drought. Comments Des Marais, “collaboration with Caroline is transformational. The methods that she developed through the J-WAFS project are changing the way I think about how to tackle my research questions.” 
Another junior faculty member from civil and environmental engineering, Benedetto Marelli, has found a research collaboration enabled by a J-WAFS seed grant impactful. Marelli is collaborating with A. John Hart, a professor in the Department of Mechanical Engineering. The two are developing an edible, silk-based food safety sensor that changes color when exposed to bacteria. “Teaming up with a senior faculty member is a good way for junior faculty members to let others know what you are doing,” says Marelli. What is more, for him the experience has involved a lot of mentoring. “Working with John [on the project], I was able to see how a really developed lab operates. As a junior faculty member, you need to immediately learn about finances, mentoring, teaching, and advising. It’s overwhelming.” Working so closely with Hart on their seed grant project, Marelli has learned from his example and shortened his own learning curve.
These few examples of how J-WAFS seed grants have influenced junior faculty at MIT provide a snapshot of the range of water and food systems research topics being pursued across the Institute. The catalyzing potential of J-WAFS seed grants not only supports these faculty members’ career advancement, but also helps to push the boundaries of water and food systems research overall. In Joann de Zegher’s words, seed grant-funded research is “early-stage research — you don’t know how it’s going to play out.” In order to go after some of the most challenging problems in water and food systems, “you need that freedom and flexibility.” More

In Song Kim, associate professor of political science at MIT, has been awarded the 2021 Levitan Prize. This award, presented each year by a faculty committee, empowers a member of the School of Humanities, Arts, and Social Sciences (SHASS) faculty with funding to enable research in their field. With an award of $30,000, this annual prize continues to support substantial projects.The 2021 award was announced by Melissa Nobles, the Kenan Sahin Dean of SHASS, who writes, “Professor Kim’s work in past years has significantly broadened access to and perspectives on social science data, helping to make legislative and policy-making processes more transparent to citizens. I am excited to see how his new project furthers the work of strengthening democracy.”
A new theory of trade
The Levitan Prize will enable Kim to move forward with an ambitious book project entitled “Trade Politics with Firms and Products.” The book “focuses on the increasing granularity in trade policies around the globe” and how that change has “fundamentally changed domestic and international trade politics.”“Representation is one of the fundamental pillars of democracy,” writes Kim, “ensuring citizens equal rights to influence public policy-making processes. However, policy-making both at home and abroad is dominated by individuals with substantial economic resources and by large multinational firms.”
Kim plans to introduce a new theory of trade liberalization that explains how “product differentiation in economic markets leads to firm-level lobbying in political markets.” In his research and in the “Trade Politics” project, Kim will use the Levitan Prize in part to collect and analyze more than 5.7 billion observations of trade and policy data from over the past three decades. Additionally, he plans to analyze 75 million lobbying reports and campaign contributions in the past 20 years.
Open-source software and classroom impact“To the best of my knowledge,” Kim adds, “this is the first book that provides a ‘big data’ analysis of contemporary trade policymaking, facilitating not only academic research of trade with a new unit of analysis but also public awareness of product-specific trade negotiations, such as the current China-U.S. trade dispute.”In addition to the book, Kim will also be developing open-source software to facilitate access to the data he will be using. The project, exploring massive datasets describing fine-grained political activities, will have a role inside MIT classrooms as well: the associated research study aims to train MIT SHASS undergraduate and graduate students to meet the interdisciplinary challenges that the project’s objectives demand. For example, Kim teaches how computational methods can be used to understand various social phenomena and policymaking in his undergraduate class 17.835 (Machine Learning and Data Science in Politics).
Making it simple to follow the path of money in politics
Kim arrived at MIT as an assistant professor of political science in 2014, directly after earning his PhD in politics at Princeton University. In 2016, he began to work as a faculty affiliate with the Center for Statistics within the MIT Institute for Data, Systems, and Society (IDSS).Kim has worked in recent years to use his research to demystify the complex financial webs of Washington lobbying and to make those connections both visible and comprehensible to the public at large. In 2018, he launched the massive public database LobbyView.org. The site draws on over 1 million public records of congressional lobbying, delivering results through a streamlined interface that vastly reduces the barrier of access to lobbying information that is, by law, already public.For instance, simply typing “Apple” into the search bar brings up a detailed history of the company’s lobbying around corporate taxes and foreign regulations. In January 2020, LobbyView was awarded the International Political Economy Society’s best new dataset award.The results of Kim’s database work, which will be advanced by the Levitan Prize, along with its web interface, will be made fully accessible to the public, to enable other researchers to interact with and export data to fuel their own work into the future.  
Story prepared by MIT SHASS CommunicationsEditorial team: Alison Lanier and Emily Hiestand More

When the Covid-19 pandemic struck in early 2020, doctors and researchers rushed to find effective treatments. There was little time to spare. “Making new drugs takes forever,” says Caroline Uhler, a computational biologist in MIT’s Department of Electrical Engineering and Computer Science and the Institute for Data, Systems and Society, and an associate member of the Broad Institute of MIT and Harvard. “Really, the only expedient option is to repurpose existing drugs.”
Uhler’s team has now developed a machine learning-based approach to identify drugs already on the market that could potentially be repurposed to fight Covid-19, particularly in the elderly. The system accounts for changes in gene expression in lung cells caused by both the disease and aging. That combination could allow medical experts to more quickly seek drugs for clinical testing in elderly patients, who tend to experience more severe symptoms. The researchers pinpointed the protein RIPK1 as a promising target for Covid-19 drugs, and they identified three approved drugs that act on the expression of RIPK1.
The research appears today in the journal Nature Communications. Co-authors include MIT PhD students Anastasiya Belyaeva, Adityanarayanan Radhakrishnan, Chandler Squires, and Karren Dai Yang, as well as PhD student Louis Cammarata of Harvard University and long-term collaborator G.V. Shivashankar of ETH Zurich in Switzerland.
Early in the pandemic, it grew clear that Covid-19 harmed older patients more than younger ones, on average. Uhler’s team wondered why. “The prevalent hypothesis is the aging immune system,” she says. But Uhler and Shivashankar suggested an additional factor: “One of the main changes in the lung that happens through aging is that it becomes stiffer.”
The stiffening lung tissue shows different patterns of gene expression than in younger people, even in response to the same signal. “Earlier work by the Shivashankar lab showed that if you stimulate cells on a stiffer substrate with a cytokine, similar to what the virus does, they actually turn on different genes,” says Uhler. “So, that motivated this hypothesis. We need to look at aging together with SARS-CoV-2 — what are the genes at the intersection of these two pathways?” To select approved drugs that might act on these pathways, the team turned to big data and artificial intelligence.
The researchers zeroed in on the most promising drug repurposing candidates in three broad steps. First, they generated a large list of possible drugs using a machine-learning technique called an autoencoder. Next, they mapped the network of genes and proteins involved in both aging and SARS-CoV-2 infection. Finally, they used statistical algorithms to understand causality in that network, allowing them to pinpoint “upstream” genes that caused cascading effects throughout the network. In principle, drugs targeting those upstream genes and proteins should be promising candidates for clinical trials.
To generate an initial list of potential drugs, the team’s autoencoder relied on two key datasets of gene expression patterns. One dataset showed how expression in various cell types responded to a range of drugs already on the market, and the other showed how expression responded to infection with SARS-CoV-2. The autoencoder scoured the datasets to highlight drugs whose impacts on gene expression appeared to counteract the effects of SARS-CoV-2. “This application of autoencoders was challenging and required foundational insights into the working of these neural networks, which we developed in a paper recently published in PNAS,” notes Radhakrishnan.
Next, the researchers narrowed the list of potential drugs by homing in on key genetic pathways. They mapped the interactions of proteins involved in the aging and Sars-CoV-2 infection pathways. Then they identified areas of overlap among the two maps. That effort pinpointed the precise gene expression network that a drug would need to target to combat Covid-19 in elderly patients.
“At this point, we had an undirected network,” says Belyaeva, meaning the researchers had yet to identify which genes and proteins were “upstream” (i.e. they have cascading effects on the expression of other genes) and which were “downstream” (i.e. their expression is altered by prior changes in the network). An ideal drug candidate would target the genes at the upstream end of the network to minimize the impacts of infection.
“We want to identify a drug that has an effect on all of these differentially expressed genes downstream,” says Belyaeva. So the team used algorithms that infer causality in interacting systems to turn their undirected network into a causal network. The final causal network identified RIPK1 as a target gene/protein for potential Covid-19 drugs, since it has numerous downstream effects. The researchers identified a list of the approved drugs that act on RIPK1 and may have potential to treat Covid-19. Previously these drugs have been approved for the use in cancer. Other drugs that were also identified, including ribavirin and quinapril, are already in clinical trials for Covid-19.
Uhler plans to share the team’s findings with pharmaceutical companies. She emphasizes that before any of the drugs they identified can be approved for repurposed use in elderly Covid-19 patients, clinical testing is needed to determine efficacy. While this particular study focused on Covid-19, the researchers say their framework is extendable. “I’m really excited that this platform can be more generally applied to other infections or diseases,” says Belyaeva. Radhakrishnan emphasizes the importance of gathering information on how various diseases impact gene expression. “The more data we have in this space, the better this could work,” he says.
This research was supported, in part, by the Office of Naval Research, the National Science Foundation, the Simons Foundation, IBM, and the MIT Jameel Clinic for Machine Learning and Health. More

At first, it seemed like the algorithm wasn’t working right.
Michael Fleder, an MIT researcher and recent alumnus working with the Laboratory for Information and Decision Systems (LIDS), had been working on an algorithm that could break down anonymized bill totals into individual item costs, creating an overview of how many people are buying a specific item or service. He was testing it out on a bulk set of data from Netflix, and although most of the data points matched to a list of the usual subscription services, there was an outlier that kept popping up at a price point too high for anything Netflix was offering.
On closer examination, Fleder realized that the algorithm was working better than expected — not only had it found known services, but it had also discovered an unannounced-but-rumored Ultra HD subscription that Netflix was testing on a limited audience. It also discovered another as-yet unmentioned product at an even higher price point.
The algorithm is detailed in a paper published at the ACM Sigmetrics Conference in December 2020 under the playful title “I Know What You Bought At Chipotle for $9.81 by Solving A Linear Inverse Problem.” Fleder co-wrote it with Professor Devavrat Shah of the MIT Department of Electrical Engineering and Computer Science, and it will be featured as part of an upcoming book by the Cambridge University Press.
Although “big data” is currently the more popular term for dealing with large amounts of information, Fleder says, “We live in this small-data problem. How can you rip these numbers apart and extract as much as you can?”
The novel inference algorithm Fleder and Shah have developed is robust, iterative, and computationally efficient, deconstructing transaction totals into the underlying products purchased, using aggregates of what is generally called “exhaust data,” or readily available anonymized data created during digital transactions.
“What is a little surprising is how the data has a signature,” says Shah. “Each individual purchase is just one number, but if many people purchase things, there is a power in collectiveness with a bit of variation, which is remarkable.”
This algorithm could be used to track sales numbers on a weekly or even daily basis, automating elements of work currently performed by financial analysts. Companies such as Google already use studies of anonymized credit data with relation to advertising, but with more detailed information readily available and increased transparency, new market opportunities may arise.
Of particular practical interest to businesses could be the increased ability to understand demand at different points of supply chains. In the case of Chipotle, their suppliers might anticipate changes in demand for ingredients like avocados by monitoring the sales of items like guacamole.
Businesses would also have new methods by which to understand and anticipate their competitors’ strategies, and it could help with businesses such as hedge funds that use transaction data to track public companies.
In its initial development, the algorithm was used on commercially-available data, provided by data vendor Second Measure. Using transactions data related to spending at Chipotle, Apple, Spotify, and Netflix, the method correctly identified the timing of the launch of a new product tier from Spotify and the release of the new iPhone XS Max. 
Fleder intends to use this algorithm as part of a new startup, with potential applications for a wide variety of companies. More

Assistant professor of biology Lindsay Case wants to understand the protein complexes called focal adhesions that let cells move and sense the world around them. She also aims to determine how cancer arises when focal adhesions malfunction. During her postdoc work at the University of Texas Southwestern Medical Center, she discovered that some of the proteins in focal adhesions work together because of phase separation — a clumping behavior that researchers are just beginning to understand. She sat down to discuss what her work means for cancer research, and her future plans for her new lab in the Department of Biology.
Q: What is phase separation, and how does it affect the way cells function?
A: I always compare phase separation to the separation of oil and vinegar. Something similar can happen with almost any kind of molecule, including proteins. When the interactions between proteins are stronger than their interactions with their surroundings, they can segregate into something called a liquid phase, similar to oil droplets floating in vinegar.
Phase separation matters because organization is a major challenge for our cells, which contain tons of different types of molecules. Sometimes cells organize these molecules using membranes, which is like using fences to keep them in place. But many subcellular structures aren’t surrounded by membranes, and how these compartments keep their components together has been a big mystery since scientists first observed them under microscopes in the 1800s. It’s really only in the last 10 years that people have realized that phase separation is part of the answer.
When cells lose the ability to stay organized, it can have devastating consequences. Changes in how proteins phase separate might underlie serious diseases, like Alzheimer’s and ALS [amyotrophic lateral sclerosis]. I’m really interested in how phase separation organizes protein complexes called focal adhesions, which link cells to their external environment. One function of focal adhesion is to let cells sense mechanical forces from the outside world, and when cells lose this ability it can contribute to cancer progression.
Q: How did phase separation initially pique your interest, and how has your research career prepared you for the work you’ll do at MIT?
A: During my PhD, I was studying how molecules within focal adhesions are organized. I saw a talk by Michael Rosen from the University of Texas Southwestern Medical Center, who would later become my postdoc advisor. Phase separation of proteins was a new idea at the time, but Mike thought it was a powerful force underlying protein organization that we needed to understand more thoroughly. I was intrigued because, at the time, I was unsure what drove focal adhesions to assemble on the plasma membrane, and I wondered if that arrangement might be due to phase separation.
I ended up joining Mike’s lab for my postdoc so that I could apply his ideas about phase separation to my interest in cell signaling and focal adhesions. As a result, I ended up working in a field as it was being born. The first year of my postdoc there were only a few papers investigating phase separation in cellular organization, and now there are over 1,000. Seeing this rapid progress firsthand has been exciting. One of the highlights of my postdoc was showing that phase separation can actually affect the functions of signaling proteins organized on membranes, and I think this discovery went a long way towards showing that phase separation isn’t just a thing that cells can do — it’s something they need to do to survive.
MIT will be an awesome place to continue studying how phase separation lets cells sense the world around them. It’s one of the institutes where the idea of phase separation in biology took off, and the MIT scientists who work on phase separation come from so many different research backgrounds. Understanding phase separation is going to require an interdisciplinary approach, which MIT values. Plus, the students are amazing!
Q: What makes your approach to studying cancer unique?
A: A lot of cancer researchers focus on large-scale or small-scale aspects of these diseases. They either look at how cancer cells behave as a whole, or study the behavior of just one protein. But there’s a level in between where I want to focus my work. I can figure out how large, multi-protein complexes like focal adhesions — some of which form because of phase separation — affect disease progression. During my postdoc, I developed a way to recreate simplified focal adhesions outside of cells. I want to use this system to learn more about how phase separation lets these complexes sense mechanical forces, and how this changes in cancer cells.
Some of the proteins found in focal adhesions are tethered to the plasma membrane, and not many people have studied how protein phase separation changes when you throw a membrane into the mix. I’m excited to keep building up my simplified focal adhesion system in my new lab, and eventually recreate the rest of the complex.
As my lab becomes more established, I’d also like to study how phase separation affects interactions between different protein complexes and signaling pathways. Phase separation is such a rapidly evolving field that it’s hard to know where my research will lead, but that’s part of the fun — not knowing where my work will take me. More