OPTIMIZEDKen-Dill(stat-madness)---Anna-Correa

Ken Dill, Ph.D., distinguished professor of physics and chemistry and director of the Laufer Center for Physical and Quantitative Biology, was one member of the Stony Brook University team in Stat Madness. ANNA CORREA/THE STATESMAN

On Feb. 27 a research team from Stony Brook Medicine showcased their skills in STAT Madness, a virtual bracket-style contest in which medical schools, universities and research institutions compete to prove that they have the best innovation in science and medicine.

Alberto Perez, Ken A. Dill, Joseph A. Morrone and Emiliano Brini represented the team from Stony Brook’s Laufer Center for Physical and Quantitative Biology. Dill serves as the director of the Center. The group of researchers entered their article, “Blind protein structure prediction using accelerated free-energy simulations,” published in the American Association for the Advancement of Science publication, Science Advances, on Nov. 11, 2016.  

It’s awesome that the research done at Laufer Center has been recognized among the best in life science innovations,” Nancy Rohring, the administrative director of the Laufer Center, said. “It says a lot about the high caliber research done here at Laufer Center and SBU.”

The Stony Brook team’s research has brought about a method involving pharmaceutical applications. The method gives a way to predict a protein structure needed to create a rational drug design.

It predicts this structure in a way that bioinformatics cannot, using a technique called Modeling Employing Limited Data (MELD). MELD uses a Bayesian inference, which is a mathematical approach that helps scientists avoid anything that isn’t a real protein, when searching through the body. It is necessary to find these real proteins so that scientists can predict their structure.

“We can’t just wait for the computer to try all possibilities – it would take longer than the age of the universe,” Perez, a Ph.D. recipient and junior fellow at the Laufer Center, said. “So instead, we use what we know about proteins, they’re compact, make secondary structure and hydrophobic residues are likely to interact with each other, and this helps us hop from place to place avoiding many possible shapes that wouldn’t look like real proteins.”

The team’s findings took them to the second round of the competition, where Stony Brook lost to Weill Cornell Medicine’s team on March 6. Cornell’s research team authored “Cell Chemical Biology A Data-Driven Approach to Predicting Successes and Failures of Clinical Trials.” The article explores a computational approach that predicts the failures of clinical trials.

Emiliano Brini ANNA CORREA/THE STATESMAN

Emiliano Brini, Ph.D., was another member of the Stony Brook STAT Madness team. ANNA CORREA/THE STATESMAN

“We made it to the round of 32 contestants, our Twitter skills were lacking,” Perez said. “In particular, we did not do a good job to engage the SBU community in knowing about this event and knowing that they could vote to help the science being done in their university to move forward.”

The competition is broken into two parts. First, a “fan-favorite winner” is chosen by online users following the competition, and then editors can pick the winner from the original contestant pool, Megha Satyanarayana, news editor of STAT, said.

We had more than one hundred entries we went through and picked,” she said. “We looked for the most innovated, what affected a large number of people, and what solved problems in life sciences – it didn’t necessarily have to be all those things but that’s what we were looking for.”

Although Stony Brook’s team did not win, Perez says that the team’s opportunity to compete in STAT Madness was unexpected yet rewarding.

“We had hit a milestone with our research that got accepted in a good visibility journal and we contacted the Media Relations Office at Stony Brook, Lauren Sheprow and Gregory Filiano in case they saw a good way to publicize it,” he said. “Greg found out about the STAT Madness competition and he thought it was worth a shot, and he certainly was right.”

Perez and his team members were initially inspired to find a new way to predict protein structure, because the current approaches used could not predict structures beyond those that were similar to the 100,000 known ones, he said.

“But simulations are lengthy, costly and have not been successful in the past, and success here would be huge for pharmaceutical applications,” Perez added.

The team says it will continue to find ways to improve their technology and make it more useful to researchers and the general public.

“We are always on the lookout for collaborations with pharmaceuticals where this kind of basic research can end up making a difference in the life of people through better health,” Perez said.

STAT Madness will announce the fan-favorite winner on April 3, and then announce the editor’s pick shortly after.

The fan-favorite winner will receive publicity on STAT’s website, and the editor’s pick will receive “bragging rights” and an article published by STAT, Satyanarayana said.

Correction: A previous version of this article erroneously misidentified Alberto Perez, a Ph.D. recipient and junior fellow at the Laufer Center and misstated the name of the publication STAT as STAT News.