Notícia (Índia)

Novel AI-based platform diagnoses Zika with 95% accuracy

Publicado em 13 junho 2018

Brazilian scientists have developed an artificial intelligence (AI) platform that can diagnose Zika viruses and other pathogens in the blood of patients with an accuracy of more than 95%. The method combines mass spectrometry, which makes it possible to identify tens of thousands of molecules that are present in the blood serum, including lipids, peptides, and fragments of DNA and RNA, using an AI algorithm that can find models related to viruses, bacteria, fungi and even of genetic origin.

A mass spectrometer is a device that acts as a sort of molecular scale, classifying molecules according to their mass. “We used the Zika virus infection as a model for the development of the platform and showed that the diagnostic accuracy exceeds 95%.” One of the main advantages is that the method does not lose sensitivity even if the virus mutates. ” Lead author Rodrigo Ramos Catharino, a professor at the University of Campinas (UNICAMP) in Brazil.

In addition, the platform, according to the study published in Frontiers in Bioengineering and Biotechnology, is able to identify positive cases of Zika itself in blood serum 30 days after the onset of infection, when the acute stage disease is over. None of the currently available diagnostic kits has the sensitivity to detect Zika infection after the end of the acute phase.

For example, the method we developed might be useful for the analysis of blood transfusion pockets, “Catharino said for the study, in 82 patients Zika was diagnosed according to the currently used method – polymerase chain reaction time – Real (RT-PCR) – recognizes viral RNA in In the 121 patients in the control group, about half had the same Zika symptoms as fever, joint pain, conjunctivitis and rash, but negative RT-PCR results: the results established a group of 42 biomarkers as a specific key to identify the Zika virus, these were in the blood of patients who were tested positive for the disease according to the researchers.