Enabling multiple cellular enumeration applications of a bioparticle sensing platform using machine learning.

Area of Science:

• Biomedical Engineering
• Computational Biology
• Analytical Chemistry

Background:

• Cellular surface receptors are vital biomarkers for diagnosing infectious diseases.
• Current diagnostic methods like flow cytometry are costly and have limitations.
• Electrically sensitive microparticles offer a novel approach for cell surface receptor detection.

Purpose of the Study:

• To improve the detection and quantification of cell surface receptors using machine learning.
• To enhance the diagnostic capabilities of impedance flow cytometry.
• To develop a next-generation cytometry technology for disease diagnosis.

Main Methods:

• Utilized microfluidic impedance flow cytometry with metal oxide-coated microparticles.
• Conjugated microparticles to blood cells targeting CD11b and CD66b surface receptors.
• Applied computational pruning and machine learning for data analysis and model training.

Main Results:

• Achieved high classification accuracies (up to 97%) after outlier removal.
• Demonstrated improved performance of neural networks with impedance spectrometry data.
• Showcased efficient biomarker quantification using machine learning models and noise reduction techniques.

Conclusions:

• Machine learning and noise reduction techniques significantly improve impedance cytometry data analysis.
• This approach enables efficient quantification of multiple biomarkers.
• The study paves the way for advanced cytometry technologies and improved disease diagnostics.