Related Experiment Video
Updated: Nov 23, 2025
08:26
Large-Scale SARS-CoV-2 Testing Utilizing Saliva and Transposition Sample Pooling
Published on: June 23, 2022
1.9K
Anomaly Identification during Polymerase Chain Reaction for Detecting SARS-CoV-2 Using Artificial Intelligence
Reynaldo Villarreal-González1, Antonio J Acosta-Hoyos2, Jaime A Garzon-Ochoa1
1MacondoLab, Universidad Simón Bolívar, Barranquilla 080002, Colombia.
Molecules (Basel, Switzerland)
|December 30, 2020
Summary
Artificial intelligence (AI) models, trained on simulated data, can rapidly verify real-time reverse transcription (RT) PCR results for Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) detection, reducing false positives and aiding diagnosis.
Area of Science:
- Molecular Biology
- Computational Biology
- Infectious Disease Diagnostics
Background:
- Real-time reverse transcription PCR (RT-PCR) is crucial for SARS-CoV-2 detection but faces verification challenges due to a shortage of trained personnel.
- Artificial intelligence (AI) offers a potential solution for automating the analysis of complex PCR data, identifying anomalies caused by contamination or artifacts.
Purpose of the Study:
- To develop and validate machine learning (ML) models for the automated verification of real-time RT-PCR data used in SARS-CoV-2 diagnostics.
- To assess the efficacy of AI in classifying PCR amplification curves and correlating results with patient data.
Main Methods:
- Generated simulated real-time RT-PCR curves across four classes: positive, early, no, and abnormal amplifications.
- Developed and tested ML models using limited simulated data, selecting the best performing model for large-scale real-time RT-PCR data from SARS-CoV-2 testing.
- Implemented a binary classification strategy, with a secondary model to differentiate negative from abnormal results, and integrated the AI into software for correlating patient and diagnostic data.
Main Results:
- Machine learning models can be effectively trained using minimal simulated data, demonstrating the feasibility of AI in molecular diagnostics.
- The AI approach successfully classified real-time RT-PCR curves, enabling rapid diagnosis and reducing the incidence of false positives in SARS-CoV-2 testing.
- The developed AI system correlated patient data with test results and AI diagnoses, showing potential for broader applications in molecular analyses.
Conclusions:
- AI, particularly ML models trained on simulated data, provides a robust and efficient method for verifying real-time RT-PCR results in SARS-CoV-2 diagnostics.
- This AI-driven approach optimizes PCR testing by enabling faster diagnoses and minimizing errors, addressing a critical need in public health.
- The methodology is adaptable for other molecular analysis applications, highlighting the versatility of AI in biological and medical research.
