Scientists at McGill University (Canada) have developed an artificial intelligence-based system that can detect incomprehensible molecular markers in the disease in individual cells. The results were published in Nature Communications (Natcom).
Scientists have explained that many diseases leave their "fingerprint" in the form of delicate changes in RNA. These signals can show the existence of the disease, its stage or the body's response to treatment. However, traditional methods often simplify data. They analyze RNA at the whole gene, skipping important information about how gene is active in each individual cell.
The processing of McGill's team allows you to analyze more details on the expression of RNA at the exhibition level, small parts of RNA, which make up the final gene structure. This approach gives a more accurate picture of molecular changes in disease.
"The genes are not built on both solid blocks, but like pieces, like Lego bricks." "Our system, called Dolphin, analyzes how these components are connected and identifying the disease signals, which were previously imperceptible," explains the first author of the study Kylu Song.
During one attempt, Dolphin analyzed the data of individual cells of pancreatic cancer patients and detect more than 800 new diseases that were not found in traditional methods. The system has also been able to distinguish between patients with aggressive and less dangerous forms of tumors.
According to the authors, Dolphin will eventually create "virtual cells". These digital models will replace real cell behavior and their reaction to medicines. This will allow us to predict the effectiveness of treatment before the clinical trials begin.
Translation of: Euromedia24.com
