Machine learning-based molecular detection using dark-field observation of two different nanoparticles

Area of Science:

• Nanotechnology
• Biomedical Engineering
• Machine Learning

Background:

• Metal nanoparticles like gold nanoparticles (AuNPs) are crucial for biosensing.
• Dark-field microscopy (DFM) analyzes AuNP aggregation for molecular detection.
• Limitations exist in detecting large molecules due to insufficient surface plasmon resonance effects at large inter-particle distances.

Purpose of the Study:

• To develop a machine learning-based method for distinguishing target-induced dimers from monomers using DFM.
• To enhance the sensitive detection of large molecules, such as proteins, by overcoming limitations of previous methods.
• To utilize dual-colored nanoparticles for improved molecular sensing.

Main Methods:

• A machine learning approach was developed to analyze single-cluster DFM images.
• Two distinct nanoparticle types (Protein A-modified silver nanoparticles and BSA-modified gold nanourchins) were used.
• Heterodimer formation was induced by the presence of anti-BSA antibody, with colors analyzed by DFM and classified using machine learning.

Main Results:

• The developed method successfully distinguished target-induced heterodimers from monomers.
• Heterodimer formation showed a concentration-dependent increase with target presence.
• The method demonstrated discrimination against non-specific aggregates and impurities like dust.

Conclusions:

• The novel machine learning-based DFM assay enables sensitive detection of large molecules like proteins.
• Utilizing heterodimer formation with dual-colored nanoparticles enhances biosensing capabilities.
• This approach offers a robust method for molecular detection, distinguishing specific interactions from background noise.