Related Experiment Video
Updated: Oct 26, 2025
06:19
Constructing and Visualizing Models using Mime-based Machine-learning Framework
Published on: July 22, 2025
1.2K
Identifying COVID-19-Specific Transcriptomic Biomarkers with Machine Learning Methods
Lei Chen1,2, Zhandong Li3, Tao Zeng4
1School of Life Sciences, Shanghai University, shanghai 200444, China.
Biomed Research International
|July 26, 2021
Summary
Accurate COVID-19 diagnosis is challenging due to similar symptoms and asymptomatic spread. This study identifies specific host biomarkers using machine learning to improve COVID-19 detection and differentiate it from other respiratory infections.
Area of Science:
- Biotechnology
- Bioinformatics
- Genomics
Background:
- Coronavirus disease 2019 (COVID-19) poses diagnostic challenges due to overlapping symptoms with other respiratory infections and asymptomatic viral transmission.
- Current diagnostic methods struggle to differentiate COVID-19 from other infectious diseases, necessitating the discovery of novel, specific biomarkers.
- The need for accurate, large-scale screening and diagnosis of COVID-19 is critical for effective public health control.
Purpose of the Study:
- To identify specific host biomarkers for COVID-19 infection using transcriptomic data.
- To develop optimized machine learning models for accurate COVID-19 diagnosis.
- To validate the ability of identified biomarkers to distinguish COVID-19 from other respiratory infections.
Main Methods:
- Utilized a publicly released transcriptomic dataset encompassing healthy controls and patients with bacterial infection, influenza, COVID-19, and other coronaviruses.
- Applied Boruta, Max-Relevance, and Min-Redundancy feature selection methods to generate an initial feature list.
- Employed incremental feature selection with classification algorithms to extract essential biomarkers and construct efficient classifiers and classification rules.
Main Results:
- Identified specific qualitative host biomarkers associated with COVID-19 infection from the transcriptomic dataset.
- Developed machine learning models capable of classifying COVID-19.
- Validated the capacity of the identified biomarkers to distinguish COVID-19 from other similar respiratory infectious diseases at the transcriptomic level.
Conclusions:
- The identified host biomarkers show potential for improving the efficacy and accuracy of COVID-19 diagnosis.
- Machine learning models applied to transcriptomic data can effectively differentiate COVID-19 from other respiratory infections.
- This approach offers a promising strategy for developing more specific diagnostic tools for COVID-19.
