Princeton University’s Top Scientific Research Of 2017: A Year In Review

Princeton University was responsible for some substantial research throughout the year of 2017. Princeton’s research in 2017 included establishing that anonymized web browsing isn’t as anonymous as people think, that air pollution in China reduces the efficiency of solar energy, and that bioinformatics can find new ways of treating pancreatic cancer. Princeton’s work throughout 2017 stands out, and while the university will no doubt go on to do great work in 2018, it is worth reflecting on all the knowledge that the university provided us with during this past year.

Anonymized Web Browsing Isn’t So Anonymous

Researchers at Princeton, in collaboration with researchers from Stanford, found that an individual’s browsing history could be correlated correctly linked to a person’s social media profile, like Facebook or Twitter.

The results were determined by building a de-anonymizer program, which took publicly available information from social media profiles and compared them with records of browsing activity. The strongest available indicators of identity were links to other websites which appeared on a social media profile, with more common links increasing the chance that a set of browsing activity belonged to a specific person. The algorithm used isn’t perfect, yet it was still able to correctly deduce, about 50% of the time, the identity of a Twitter profile when examining 20 links originating from Twitter.

This means that while large companies like Google and Facebook track their users, so could anyone with access to a person’s browser history, which includes a substantial amount of different corporate entities and organizations. Yet unlike Facebook or Google, these private trackers may not choose to disclose the fact that they are tracking a person’s activity. Furthermore, provisions like the FCC’s internet privacy rules state that a company may only log data when the information is not “reasonably linkable” to individuals. The new study shows that standards for “reasonably linkable” may have to be redefined.

Premature Babies and Sensory Brain Processes

Research done by Princeton scientists has found that babies who are born prematurely don’t use expectations about the world to influence the development of their brains like normal babies do.

At-term babies use a portion of their brain responsible for processing visual stimuli, and it responds to what the baby expects to see, not only what they do see. This means they are learning from experience. Yet pre-term babies don’t use this system.

The results were determined from a series of tests run on 100 babies, six months of age, some of which were born at-term and some of which were born prematurely. The babies were shown images and sounds, coupled with each other to produce predictable image/sound pairs. The researchers then began removing the image and only playing the sound. Only the full-term babies displayed activity in the visual system, which suggests that they are predicting that the image should appear. The pre-term babies did not show this brain activity.

The researchers hope that the study will guide scientists in determining what goes wrong in the brain in cases of prematurity.

Air Pollution Reduces Solar Energy Output

China is one of the largest investors in solar power in the world. The nation is hoping to meet 10% of its electricity needs with only solar power by 2030. Complicating this matter is that the output of energy from their solar panels is being substantially reduced by air pollution, according to a study published by Princeton researchers.

The pollution which blocks out the sun in China comes from the combustion of organic compounds, either the burning of biomass for cooking or heat or the burning of fossil fuels in power plants and in cars. During the winter, when the most fuel is burned for heating, air pollution can block almost 20% of sunlight from reaching the solar cells in solar panels. These calculations were achieved by combining satellite data collected on pollution levels with a solar panel performance model. Nine different analyses were done between 2003 to 2014.

The pollution is the worst in northern and eastern China, where it reduces energy generation by around 35%. That’s equivalent to approximately half a kilowatt-hour a day, enough to run a laptop for between 5 to 10 hours. It’s hoped that the findings will encourage the Chinese government to pursue the reduction of aerosol emissions, in order to improve the efficiency of the solar panels they are investing so heavily in.

Model Created For Genetic Regulation of 44 Human Tissue Types

We know that our genes determine who we are, how healthy we are, and what functions our bodies carry out. Researchers from Princeton managed to make substantial progress towards creating a gene-based model that accounts for the health of bodies as well as how certain diseases develop. The Princeton researchers collaborated with researchers from around the world to deliver a report detailing how genetic variation impacts gene regulation for 44 different types of human tissue.

The research team obtained different tissue samples, including samples of organs, blood, and brain material from donors immediately after their death. They analyzed over 7000 tissue samples to investigate how genes are expressed and regulated within various types of tissues. These rare tissue samples allowed the researchers to characterize the function of a wide range of healthy tissues, and use them as a baseline to compare how disease alters the function of these tissues. Thanks to the wide variety of samples, the largest such study to date, scientists now have a chance to examine the effects of genotypes which heighten an individual’s risk for certain diseases.

A Bioinformatics Focused Treatment For Pancreatic Cancer

Pancreatic cancer is the third deadliest disease in the nation. A major computational analysis of pancreatic tumors may have paved the way for new forms of treatment for the disease.

Researchers at Princeton did a computational analysis of pancreatic tumors samples from 150 patients, looking for differences that occurred between cancerous pancreatic cells and normal cells. The results of the study implied that a wider cellular communication pathway, as well as a few specific genes, played causal roles in the development of the disease. A gene dubbed KRAS was found to be mutated in the almost all (93 percent) of patients within the study. Even amongst the seven percent of patients who lacked this genetic mutation, 60% of them had mutations in genes which were part of KRAS’ same signaling pathway.

Signal pathways are responsible for passing information to the nucleus of a cell from outside of the cell. Correcting anomalies in this signaling pathway, such as a malfunctioning protein created by KRAS, could prove effective in treating not only pancreatic cancer but other types of cancer as well.

The study was also successful in identifying changes in the genome of pancreatic cancer cells which should be receptive to current treatments or treatments soon to be available, meaning that the treatment of pancreatic cancer could see substantial improvements rather soon.

Creating A Mega-Lab To Study Cognition And Behavior

Neuroscientists from Princeton are collaborating on the creation of a virtual mega-laboratory designed to study interactions between individual neurons and discern how the brain makes choices.

The new laboratory will be referred to as the International Brain Laboratory (IBL), and its goal is to instantiate a level of uniformity for experiments on a massive scale. Traditionally, different labs use different experimental procedures to get results, but this makes comparisons between experiments difficult. Those who take part in the IBL project will carry out the same experimental procedures for a specific task, enabling researchers around the world to aggregate their data as if a single massive experiment were being run.

A variety of labs will be investigating the neuronal activity of mice performing decision-making tasks. The animal will be given a visual stimulus, and have to perform some action that involves making a choice based on the stimulus. While the mice are making the choices, the team from IBL be recording the activity of all regions of the brain, monitoring between 5000 to 10,000 neurons in the animals as they perform the exact same task around the world.

New tools and technological innovations will be created to assist in the development of large-scale models based on the data collected from the many different labs. IBL researchers want to use the information gained from monitoring various parts of the brain to understand how each part of the brain contributes to decision-making, with the ultimate goal of creating a comprehensive theory of decision-making.

Trust In Community Enables Poor To Make Better Decisions

Princeton researchers discovered some interesting and important findings in the areas of psychology and economics. A study done by Princeton researchers found that low-income people make superior long-term financial decisions when they are able to trust in their local communities. If people are able to rely on neighbors and friends for financial support, they are less likely to turn to solutions such as payday loans.

The findings were established by a series of studies that examined the relationship between short-term financial decisions and low-income people. The first study examined how people across different economic levels made financial decisions, finding that more well-off participants in the study were less likely to make short-term decisions that harmed them. Low-income people were more likely to make harmful short-term decisions, but only low-income people who lacked a community that they could trust.

Supporting this study was a second study which examined the impact of unreliable communities on people’s relationship with payday loans. Those who have friends, neighbors, or other community members they can borrow money from in an emergency take out fewer payday loans (loans with high interest, which typically exacerbate cycles of poverty).

Finally, a third study found that people in Bangladesh were more likely to trust in their community if their community had representatives trained by NGO’s. This also impacted their decision-making, and those who felt they could approach community representatives were more likely to make smart financial decisions in the long run.

Overall, the research points to the fact that investing in community development programs, particularly those which develop community trust and involvement, can help alleviate poverty.

These are some of Princeton’s most notable accomplishments during 2017, and more impressive accomplishments are sure to come this year.