Freie Universität Berlin

Working in close collaboration with other institutions, researchers at this university used a massively improved computing environment to create a solution that enables rapid identification of certain cancers so patients can be alerted and treated earlier



Freie Universität Berlin is a leading research institution. It is one of nine German universities successful in all three funding lines in the federal and state Excellence Initiative, thereby receiving additional funding for its institutional future development strategy. The university’s performance in the Excellence Initiative has provided funding for several new graduate schools and transdisciplinary research clusters. Freie Universität Berlin is a member of the IBM Academic Initiative.

Business need:
Researchers at Freie Universität Berlin and MATHEON, both world-class German research institutions, sought better methods for analyzing human blood proteins as a means of detecting diseases earlier than was previously possible. Their focus: cancer. Their general approach: proteomic pattern diagnostics, a relatively new method for the early detection, surveillance and monitoring of diseases through the identification of disease-specific proteomic “fingerprints”—roughly, specific, abnormal mixtures of proteins in the blood.

Solution:
By working closely with mathematicians and other collaborators, computer scientists at the university created a solution enabling scientists to analyze 12,000 patient records in real time—each comprising roughly 2.5 GB of data—at a rate approximately 250 times faster than they had before. The solution analyzes and compares samples of healthy and diseased individuals by applying statistical algorithms that help identify disease-specific features of proteome profiles. When moved to a clinical setting, doctors and their patients will get more test results and more precise diagnoses—faster.

Benefits:
-Increased the speed of analyzing patient records by 250 times -Allowed scientists to examine 12,000 2.5 GB patient records in real time -Detected and validated new “fingerprints” for bladder, kidney and pancreatic cancers—a result that, according to the university’s clinical partners, would not have been possible using current blood-testing techniques -Lent a boost to the search for earlier and less aggressive cancer therapies that have a positive impact on a patient’s quality of life and prognosis, as compared with existing treatments
.

Case Study



Freie Universität Berlin is a leading research institution. It is one of nine German universities successful in all three funding lines in the federal and state Excellence Initiative, thereby receiving additional funding for its institutional future development strategy. The university’s performance in the Excellence Initiative has provided funding for several new graduate schools and transdisciplinary research clusters. Freie Universität Berlin is a member of the IBM Academic Initiative.

The Opportunity
Researchers at Freie Universität Berlin and MATHEON, both world-class German research institutions, sought better methods for analyzing human blood proteins as a means of detecting diseases earlier than was previously possible. Their focus: cancer. Their general approach: proteomic pattern diagnostics, a relatively new method for the early detection, surveillance and monitoring of diseases through the identification of disease-specific proteomic “fingerprints”—roughly, specific, abnormal mixtures of proteins in the blood.

What Makes It Smarter
One of the primary challenges in proteomics is dealing with the immense data sets involved in comparing and mapping proteins at a speed that will help save lives. By working closely with mathematicians and other collaborators, computer scientists at the university created a solution enabling scientists to analyze 12,000 patient records in real time—each comprising roughly 2.5 GB of data—at a rate approximately 250 times faster than they had before. The solution analyzes and compares samples of healthy and diseased individuals by applying statistical algorithms that help identify disease-specific features of proteome profiles. When moved to a clinical setting, doctors and their patients will get more test results and more precise diagnoses—faster.

Real Business Results
- Increased the speed of analyzing patient records by 250 times
- Allowed scientists to examine 12,000 2.5 GB patient records in real time
- Detected and validated new “fingerprints” for bladder, kidney and pancreatic cancers—a result that, according to the university’s clinical partners, would not have been possible using current blood-testing techniques
- Lent a boost to the search for earlier and less aggressive cancer therapies that have a positive impact on a patient’s quality of life and prognosis, as compared with existing treatments