Direct biomolecule discrimination in mixed samples using nanogap-based single-molecule electrical measurement

Area of Science:

• Nanotechnology and Molecular Electronics
• Machine Learning Applications
• Analytical Chemistry

Background:

• Single-molecule measurements using metal nanogap electrodes offer direct current analysis.
• Machine learning enhances single-molecule signal identification accuracy.
• Conventional methods face limitations with varied electronic structures and per-molecule training data.

Purpose of the Study:

• To develop a molecule identification technique using single-molecule measurement data from mixed solutions only.
• To overcome the limitations of conventional methods requiring individual sample training and facing nanogap variations.
• To enable molecule identification without prior training on individual components.

Main Methods:

• Utilized single-molecule measurement data from mixed sample solutions.
• Applied machine learning algorithms to analyze current signals from single molecules.
• Developed a predictive model to determine mixing ratios from mixed-solution data.

Main Results:

• Successfully predicted the mixing ratio of molecules in solutions using only mixed-sample data.
• Demonstrated accurate molecule identification without prior training on individual samples.
• Validated the feasibility of identifying single molecules solely from mixed solution measurements.

Conclusions:

• The proposed method allows for molecule identification using only mixed solution data, bypassing traditional training requirements.
• This technique significantly advances single-molecule measurement applications, especially for biological samples where separation is difficult.
• It broadens the potential adoption of single-molecule measurements as a versatile analytical tool.