Big Data for Personalized Medicine
How "big Data" Helps genetic Research
Irina Paroshina, ComputerraDiscoveries in the field of genetics have become systematic for some time.
However, they still take a lot of time and effort. Uncovering patterns hidden in medical records and DNA databases is not an easy task. Testing a medical hypothesis can take so much valuable time that scientists unwittingly try to be more restrained in their experiments and theories. Moving away from the use of traditional IT systems, researchers have gained greater freedom of experiment and the opportunity to shorten the study period from a year to several weeks.
The Vanderbilt University School of Medicine turned to solutions based on the IBM platform. The analysis of huge amounts of information accumulated by the university over the years of its work is now performed using IBM PureData. "Big Data" has become not just an optimization of the research process, but certainly a new stage in the development of translational medicine – an interdisciplinary science combining a clinical approach with new technological methods. The principle of translational medicine is to transfer fundamental scientific research from research centers and laboratories to medical practice.
The new platform gives clinicians and researchers the opportunity to analyze clinical data and DNA data of 2.2 million patients accumulated over twenty years of the university's work. It also helps researchers and physicians to combine pheno- and genetic markers with information about the health of the population. As a result, by understanding the genetic predisposing factors of the disease, doctors can adapt care and improve treatment outcomes.
The essence of the pioneering work at Vanderbilt University is to identify the genetic nature of the disease and determine the response to drugs. Which patients are at risk of being in the group of cases and why some of them do not respond to certain types of drugs – these are the questions that scientists are asking. It is necessary to understand the causes and develop new methods of therapy and prevention of diseases. To do this, a large database called Synthetic Derivative was created. In addition, BioVU scientists have at their disposal a database of Vandebilt DNA. However, the data itself is not enough: combining specific features of geno- and phenotypic markers with specific diseases and health consequences is a task that requires the ability to observe billions of medical records in parallel in various logical slices.
The PureData system made it possible to effectively manage large-scale data. Scientists have managed to isolate about 100 billion genotypes, and comparing their queries, which used to take hours, now require several minutes. This accelerates the pace of research. Algorithms can be developed in real time – it's trivial to sit down with a colleague and get results and refine methods within a few hours. This means that you can test more theories, find ways that could not have been predicted initially. You can no longer exclude ideas that have a lower probability of being successful. Previously, these were simply dismissed in order to save time, but often they have unexpected advantages. Thus, the acceleration of the pace of research encourages scientists not to postpone experiments and noticeably diversify them.
Working with "big data" has already yielded the first results. For example, it was found that in patients with a certain genetic variance, when prescribing a specific (and, moreover, quite popular) antiplatelet drug, a new heart attack is likely to occur. Such a patient, compared with others for whom treatment is quite effective, has a high risk of dying from this attack or stroke. The variant found by PureData was tested on 13 thousand people. As a result, therapy for a number of patients has been changed. In another study, scientists were able to determine that people with certain phenotype variations are more likely to develop arrhythmias than others. This knowledge allows doctors to take into account this risk in advance and prevent an unfavorable outcome. These are just two examples, while in the near future the organization plans to study forty different diseases and twenty drugs.
So, Big Data technologies do not just reduce the research time from a whole year to several weeks. They help doctors identify the increased risks of developing new syndromes and recurrent seizures in patients and prescribe appropriate medications in time.
Portal "Eternal youth" http://vechnayamolodost.ru25.12.2013