Brazilian researchers devise AI-based approach to detect pathogens

15 June 2018 (Last Updated June 15th, 2018 11:48)

A research team from the University of Campinas (UNICAMP) in Brazil has developed a new platform based on artificial intelligence (AI) for the diagnosis of various pathogenic diseases.

Brazilian researchers devise AI-based approach to detect pathogens
The algorithm recognises patterns associated with pathogenic diseases. Credit: Wikimedia Commons.

A research team from the University of Campinas (UNICAMP) in Brazil has developed a new platform based on artificial intelligence (AI) for the diagnosis of various pathogenic diseases.

The approach uses a combination of AI algorithm and mass spectrometry to accurately identify metabolic markers in patient blood samples.

Mass spectrometry can detect numerous molecules present in blood serum, while the algorithm can pin-point patterns related to viral, fungal, bacterial and genetic diseases.

“It is expected that the algorithm can examine large data volumes for specific patterns that can help in classification, prediction, decision making and modelling.”

The team built the platform using Zika virus infection as a model and observed more than 95% accuracy for diagnosis of this condition.

According to the researchers, the technique retains sensitivity even in cases of virus mutations, with the capability to detect positive Zika cases even in samples examined after completion of acute infection phase (30 days post onset).

UNICAMP Innovare Biomarker Laboratory head Rodrigo Ramos Catharino said: “None of the currently available diagnostic kits has the sensitivity to detect infection by Zika after the end of the acute phase. The method we developed could be useful to analyze transfusion blood bags, for example.”

The researchers trained the AI tool using data obtained from analysis of 203 blood samples, of which 82 patients were diagnosed with Zika using real-time polymerase chain reaction (RT-PCR) method.

It is expected that the algorithm can examine large data volumes for specific patterns that can help in classification, prediction, decision making and modelling.

Catharino noted: “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.”

Currently, the team is assessing the platform for the diagnosis of fungal systemic diseases, with plans to extend the evaluation to bacterial and genetic conditions.