A year ago, the Icahn School of Medicine at Mount Sinai (ISMMS) took a giant step toward tackling some of the most difficult questions in science and launched a new era in scientific computing by building and operating “Minerva,” an on-site supercomputer for genomics and other basic sciences. In that short time, Minerva has helped ISMMS scientists analyze major diseases such as cancer, diabetes and Alzheimer’s with more precision, and it has also brought the medical world closer to developing more effective drugs and obtaining more accurate pathology results.
Driving Personalized Cancer Therapy
Eric Schadt, PhD, Professor and Chair for Genetics and Genomic Sciences and Director of the Icahn Institute for Genomics and Multiscale Biology, and his team rank as Minerva’s biggest users. With their research studies offering novel insights into the evolution and signaling of different cancers, they have discovered a vast diversity in most tumors, a diversity that may continue to increase as the tumors mutate to evade treatment. This information has led them to understand that, although analyzing genetic data is vital to delivering the right treatment for a patient’s cancer, it is not the only piece of the puzzle.
To take cancer therapy to an even more personalized level, Dr. Schadt’s team has launched a program that adds predictive network modeling to genetic sequencing. By combining genetic information with patient-specific mutation analysis, the team aims to identify disruptions that might be missed in targeted gene panels and pinpoint the most appropriate therapy for each individual patient.
“The emerging ability to deliver the right treatment, in the right dose, and at just the right time, especially to very sick patients, is completely revolutionizing the practice of medicine,” says Dr. Schadt.
Conquering Drug-Resistant Infections
Assistant Professor for Genetics and Genomic Sciences, Ali Bashir, PhD, a member of Dr. Schadt’s team who specializes in infectious diseases and cancer, is currently running seven projects on Minerva. He notes that we have developed a 24-hour turnaround from sequencing to high quality assembly of bacterial genomes, and he sees that as a strong weapon in combatting drug-resistant infections like MRSA. According to Dr. Bashir, “The methods we are developing for assembly and strain comparison are ideally suited for strains like MRSA. By better understanding viral and bacterial pathogens at the genomic level, we can direct current therapeutics to their best use, and we can also obtain insight into how to target new pathogens more precisely.”
Developing Safer Drugs with Fewer Side Effects
Scientific computing may be the key to developing more effective drugs with fewer side effects. Marta Filizola, PhD, Associate Professor for Structural and Chemical Biology, has a keen interest in the family of cell surface proteins called G protein-coupled receptors, which are the targets for about half the drugs currently in use. Her lab is studying how these receptors, which are never static, change form to interact with drugs and other intracellular proteins. These interactions convert into biological signals in the interior of cells, thereby creating the body’s response to disease or pain. By investigating this complex process of cell signaling, Dr. Filizola aims to create more specific and effective drugs, and, with opioid receptors factoring strongly in her research, she is seeking to develop pain medications that do not cause addiction.
Moving From Cell Biology to Systems Pathology
ISMMS is one of only two institutions in the nation to run a unique software program called Virtual Cell, or VCell. Developed by Leslie Loew, PhD, and his colleagues at the University of Connecticut, the program offers a unique framework for integrating cell biological mechanisms with imaging data to understand cell shape and morphology in health and disease.
Ravi Iyengar, PhD, Professor and Chair for Pharmacology and Systems Therapeutics and Director of the Experimental Therapeutics Institute, has recently uploaded VCell to Minerva to study of the dynamics underlying cell shape and their relationship to disease states. In a recent study he and his colleagues have found that cells with fusiform (elliptical shapes), such as those seen in many cancer tissues, enhance growth factor signaling that can lead to greater proliferation. He predicts that such mechanistic relationships between the shape of cells within tissues and molecular determinants have the potential to transform clinical decision making by unraveling a larger piece of the diagnostic puzzle. Results from his studies, when combined with a patient’s genetic information, will be able to offer physicians remarkably accurate and specific clinical information upon which they can base their treatment decisions.
According to Dr. Iyengar, “VCell functions mathematically through partial differential equations on real complex shapes observed by imaging. It is the only program in the world that works at such a complex level. It has the ability to produce sharp, clear images that scientists have never seen before.” He compares today’s brain scan to an impressionist painting. Scientists can see the picture, but it is not sharp. Images produced through VCell will be clear, similar to an actual photo, and prove infinitely more useful in diagnosis.
Running VCell on Minerva also offers a “big data” advantage. Since we can store huge amounts of information on our supercomputer, process it at an incredible speed, and generate data inter-knowledge, we will be able to provide physicians with information that will allow them to take precision medicine to a much more accurate and sophisticated level of care. He anticipates achieving this breakthrough within the next decade, possibly even sooner when analyzing brain and kidney cancers and other kidney diseases.
Analyzing Data on 150 Projects
Named after the Roman goddess of wisdom and medicine, Minerva analyzes data through sophisticated algorithms. The $3 million computer, one of the first to be housed in an academic medical center, is currently running 150 projects, and our 300 users have logged some 60,000,000 hours of operation. These scientists benefit greatly from the fact that we have recruited some of the brightest stars in science and technology to produce reliable and timely interpretations of data and analyze genomes at a faster rate than ever before.
Minerva has enjoyed a remarkable first year with our scientists producing significant results and great promise. We anticipate breakthrough discoveries from current and future users as we move forward into year two.