TMC PULSE

t m c » p u l s e | j u n e 2 0 1 4 12 T he sheer volume of data in our world has been expanding expo- nentially. Analyzing large, complex sets of data has become a necessity in understanding everything from the wealth of information derived from social media to data obtained from analysis of the human genome. From the office of the White House to every pharmaceutical and health care orga- nization in the world, data is the key strategic asset that has the potential to change the way that we provide patient care, cure diseases and save lives. In 2011, McKinsey & Company estimated that, "If U.S. health care were able to use big data creatively and effectively to drive efficiency and quality, more than $300 billion in value would be created every year." With an annual patient volume that exceeds 7 million and a vast, rich network of research enterprises, the Texas Medical Center is a prime contender for a revolutionary approach to data analysis. In keeping with their endeavors to create a world-class center for complex data research and opera- tional excellence in medicine, the medi- cal center has selected Ayasdi, a startup company from Palo Alto, California that uses topological data analysis to quickly glean meaning from big data, as part of its vision to revolutionize patient care, translational medicine and clinical research. In 2012, the health care industry amassed over 500 petabytes (a unit of data measurement) of health care data; this number is on the trajectory to grow to 25,000 petabytes by 2020. Although we have figured out how to store, manage and process data, the ability to derive insights from that information has not changed much in the last 30 years. In most cases, organizations rely on labor-intensive, hypothesis-based analytical approaches that exclude anyone without advanced degrees in computational mathematics and statistics. With exponential data growth estimates and increased data com- plexity, new data analysis techniques like Ayasdi's will be integral to finding breakthroughs and innovations from health care data. "We regard this partnership as a tremendous opportunity," said Gunnar Carlsson, Ph.D., co-founder of Ayasdi and professor of mathematics at Stanford University. "The Texas Medical Center has an enormous vari- ety of very sophisticated institutions. It provides the perfect place to use our software because of the sheer volume of data being generated. The scientific potential of their research enterprises is unparalleled." Ayasdi's software is based on topo- logical data analysis, which visualizes relationships between data points Illuminating Data topological data analysis helps identify patterns, relationships in medical data Mount Sinai utilized Ayasdi Cure—which uses machine learning and topological data analysis—to analyze over 20,000 patient medical records paired with genetic data. Because Ayasdi's approach is query-free, and can automatically map relevant data points based on hundreds of algorithms, Mount Sinai was able to gain new break- through insights that can dramatically impact how Type 2 Diabetes patients are diagnosed and treated. (Photos by: Ayasdi) as two dimensional shapes made up of interconnected nodes. The nodes represent the subjects under analysis, whether they are a dataset of breast can- cer tumors or statistics of players from the Miami Heat—the team leverages Ayasdi's software to analyze basketball data in order to make smarter match up decisions. The most similar nodes will cluster together, with lines drawn between them, while dissimilar nodes will be separated into different shapes. The idea is that emergent patterns, like loops, forks, or separated clusters, will reveal meaningfully different subpopu- lations in the dataset. "Think of studying data like look- ing for a hammer in a dark garage," explained Carlsson, who specializes in a branch of mathematics called topol- ogy, the study of shape. "If you have a flashlight and point it at the ceiling, or in the corner, it's not very likely that the narrow beam of light will illuminate your target. If you were able to turn on the light in the garage, your eyes would be drawn to the hammer and maybe a bucket of nails in the corner you hadn't noticed. The idea is to visualize the data in order to see the whole statisti- cal landscape, generating questions accordingly. It's the Jeopardy method of data analysis: answers first." Topology, the subset of mathemat- ics that focuses on the study of shapes, dates back to the 19th century, but scholars were writing treatises on the subject for over a century beforehand. While topology has existed, in a theo- retical sense, since the 1700's, Carlsson pioneered the applied use of topology to solve complex, real world problems beginning in the mid 1990's. In the early 2000's, this work led to $10 million in research grants, courtesy of the National Science Foundation (NSF) and the Defense Advanced Research Projects Agency (DARPA) to study the application of B y A l e x O r l a n d o

• Cover