Related Experiment Videos
Machine learning for the meta-analyses of microbial pathogens' volatile signatures
Susana I C J Palma1, Ana P Traguedo1, Ana R Porteira1
1UCIBIO, REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516, Caparica, Portugal.
Scientific Reports
|February 22, 2018
Summary
Researchers identified specific microbial volatile organic compounds (VOCs) using artificial intelligence. These VOCs can accurately detect 11 human pathogens, advancing non-invasive infectious disease diagnostics.
Area of Science:
- Biotechnology
- Computational Biology
- Infectious Disease Research
Background:
- Volatolomics offers promising non-invasive diagnostic tools for infectious diseases.
- Identifying microbial volatile organic compounds (VOCs) that differentiate human pathogens remains a challenge.
- Artificial intelligence (AI) is a crucial tool in modern health sciences.
Purpose of the Study:
- To develop a method for identifying microbial VOCs with pathogen-discriminating power using machine learning.
- To establish sets of VOCs capable of accurately identifying specific human pathogens.
Main Methods:
- Machine learning algorithms, specifically support vector machines and feature selection, were employed.
- Analysis of published studies (1977-2016) on VOCs emitted by human microbial pathogens.
- Development of pathogen classification methodology adaptable for future data integration.
Main Results:
- A set of 18 VOCs demonstrated high accuracy (77%) and precision (62-100%) in predicting the identity of 11 microbial pathogens.
- Specific VOC sets were identified for each pathogen, achieving high accuracy (86-90%) in predicting their presence.
- The methodology supports database expansion with new pathogen-VOC data.
Conclusions:
- Identified VOC sets can significantly improve the selectivity of non-invasive infection diagnostics.
- This approach paves the way for enhanced artificial olfaction devices for pathogen detection.
- The findings contribute to the advancement of AI-driven diagnostic tools in healthcare.