Universities of Campinas (UNICAMP) in Brazil scientists and researchers have developed a new artificial intelligence (AI) platform that can detect the Zika virus even after the 30-day acute phase of the disease. Zika detection after the acute phase has never been achieved before. The AI platform has been able to diagnose several diseases accurately by detecting the metabolic markers of the diseases in the blood.
The platform uses mass spectroscopy to find the metabolic markers. Mass spectroscopy can identify tens of thousands of molecules in the blood. The AI algorithm uses machine learning to detect patterns in the molecules that indicate what disease is present in the blood sample. The system has the potential to detect viral, bacterial, fungal and genetic diseases. The current method for detecting Zika is called real-time polymerase chain (RT-PCR). RT-PCR detects viral RNA in body fluids during the acute stages of diseases like Zika.
During the development of the AI platform, 203 blood samples that were taken a UNICAMP’s teaching and general hospital were analyzed by a mass spectrometer. 82 of these blood samples had been diagnosed by RT-NRA to have Zika. The other 121 patients were used as the control group in the study. Half of these patients had symptoms of Zika but were not diagnosed by RT-NRA and the other half has no symptoms and were tested negative for Zika by RT-NRA or were diagnosed with dengue.
“We identified some 10,000 different molecules in the patients’ serum, including lipids, peptides, and fragments of DNA and RNA. Among these metabolites, there were particles produced both by Zika and by the patient’s immune system in response to the infection,” said Rodrigo Ramos Catharino, principal investigator for the project.
The data collected by the spectrometry analysis of the positive group and the control group were then given to a computer program with a random-forest machine learning algorithm. The data allowed the AI-system to learn more and grow more, making it more sensitive to the virus to detect the disease.
“The algorithm separates samples randomly, determines which one will be the training group and the blind group, and then carries out testing and validation. At the end, it tells us whether with that number of samples it was possible to obtain a set of metabolic markers capable of identifying patients infected by Zika,” Catharino explained.
The study created 42 biomarkers for the cases of Zika. 12 of these biomarkers were found in the Zika blood samples by the algorithm.
The UNICAMP group is currently testing the platform’s ability to diagnose all kinds of systematic diseases that are caused by fungi. The goal is to have the algorithm detect many diseases, including bacterial or genetic, in both the early and late stages of the symptoms.
The researchers say that any lab with a mass spectrometer would be able to use the platform for disease diagnosis.
The paper on the program was published in Frontiers in Bioengineering and Biotechnology.