26 November 2013

About the benefits of archived data

How Old Medical Histories Help Medical Genetics

Kirill Stasevich, CompulentaA lot of diseases arise due to malfunctions in DNA - if not entirely, then at least in part, and scientists, of course, try to establish the fullest possible connection between the two.

And usually it happens like this: some kind of disease is selected and the people suffering from it - the more, the better; they analyze their DNA and compare it with the DNA of those who do not have this disease.

Such studies require a lot of time and money: you need to find a sufficient number of patients, collect information about their genomes, etc. All this can last for years, so scientists are constantly looking for a way that would save both time and money. And one of these approaches was proposed by researchers from the Vanderbilt University School of Medicine (USA).

Strictly speaking, what is reported in Nature Biotechnology (Systematic comparison of phenome-wide association study of electronic medical record data and genome-wide association study data) Joshua Denny and his colleagues seem so obvious that it's amazing how it hasn't occurred to anyone until now. In medicine (at least in Western medicine), it has long been practiced to keep electronic records reflecting the entire patient's medical history: diagnosis, treatment, test results, doctor's comments, as well as billing information required by an insurance company to ensure payment for services.

It occurred to the researchers that this information could well be used anew to identify links between genetic variants and different diseases. The work used a large array of data related to 13,835 electronic medical records: all of them were grouped into categories of diseases, numbering 1,358. These records also contained information about the DNA of patients, so it remained only to compare DNA information with diseases.

And "big data" has again borne fruit: according to the authors of the work, some of the genetic variants that were already known from past genomic studies as associated with certain ailments have quite reliably confirmed this connection. Moreover, it was possible to establish "painful" coincidences for some nucleotide variants in DNA that have not yet come to the attention of genomic studies. Among the diseases that received new (possible) genetic causes were non-melanoma skin cancer and two forms of keratosis, one of which is again considered a precursor to cancer. In an additional study, it was possible to confirm these findings once again. In general, there is clearly more work for molecular biologists who are engaged in deciphering specific mechanisms linking DNA anomalies and cellular and physiological ill health.

However, these results, for all their significance, in this case are only side effects. The main thing is that scientists have managed to prove the effectiveness of this approach, when the study does not start from scratch, but uses old medical records. Moreover, the method has one advantage: it can be used to cover many diseases at once (whereas in a conventional genomic search, researchers focus only on one disease).

At the same time, as it is easy to understand, it is necessary that these medical data be in a systematized electronic form. And although such an "electronic" way of working in some medical institutions has long become common practice, in some places it is still viewed as the last word of progress. In addition, these records should contain information on DNA, and, alas, the corresponding analyses are not always done. However, it is DNA sequencing in medical institutions that is being carried out more and more often (and it is becoming easier and cheaper to perform it), so that soon it will be impossible to complain about the lack of such materials for research.

The authors of the work themselves emphasize that the volume of processed data can be expanded further (it would be something to process them) and that such comparisons of medical histories can give a lot of information about rare but important clinical precedents (for example, about the side effect of the drug, which can manifest itself in one patient among 10 thousand).

By the way, not so long ago we reported how a fairly old and well-known clinical method is gaining a second wind thanks to "big data": it was about the fact that the good old blood test can accurately predict the risk of premature death from heart and vascular problems.

Prepared based on Technology Review: Genetic Link to Skin Cancer Found in Medical Records.

Portal "Eternal youth" http://vechnayamolodost.ru26.11.2013

Found a typo? Select it and press ctrl + enter Print version