High-Precision Single-Molecule Identification Based on Single-Molecule Information within a Noisy Matrix

Abstract

When conducting a single-molecule measurement, data are analyzed using the histogram of a measured physical quantity in which a single dataset contains information about a specific single molecule. Oftentimes, the histogram consisting of only specific single-molecule information excludes the input from other information sources. In other words, despite measuring the single molecule during analysis, we miss many of the properties of that single molecule. Herein, we have successfully identified a single molecule with a high degree of precision via a one-electric current pulse method using machine learning to read the single-molecule information. With the use of positive unlabeled classification, which is one of the techniques used in machine learning, we have developed a method for discerning a single molecule from a background of electric noises by analyzing the electric noises produced at the nanoscale level. In this method, we have demonstrated that the 2-, 3-, and 4-type nucleotides could be identified with a high degree of accuracy at a single-molecule resolution. This method can be widely applied for the accurate identification of a nucleotide using one measurement signal within a noisy matrix.

Time-resolved neurotransmitter detection in mouse brain tissue using an artificial intelligence-nanogap

“…The single neurotransmitter signals after noise removal 28 , 29 were classified via supervised ML using the XGBoost classifier 30 . We used the data with a gap width of 0.56 nm because the detection rates of the three neurotransmitters are the same (see Figure S5 c).…”

“…The current signals included various noise signals from other neurotransmitters, contaminants and the migration of the gold electrodes. These noise signals were removed using a positive and unlabelled data (PU) classifier 28 , 29 , a suitable algorithm for distinguishing between two classes when only positively labelled and unlabelled data are given. The detailed analysis procedure is described in the Supplementary Information.…”

“…Improvement of the discrimination method in single-molecule measurement is effective for investigating neurotransmitter interplay. To solve this problem, we developed an analysis method based on machine learning (ML) from the waveform of the tunnelling current flowing between nanoelectrodes via a single-molecule 28 . ML enables higher accuracy identification of multiple types of target molecules (e.g., neurotransmitters in the brain) in the presence of noise.…”

Analysis Method of the Ion Current-Time Waveform Obtained from Low Aspect Ratio Solid-state Nanopores

“…The identification of single molecules in the contaminants (noise) has been demonstrated. 106 The analytical system measures and learns on a sample that does not contain an analyte, in which case the resulting signal is labeled as noise. When a solution containing the analyte is measured, the signals obtained include both the noise and the analyte signal.…”

An artificial intelligence nanopore platform for SARS-CoV-2 virus detection

“…In clinical specimens that are enriched in contaminants, the positive specimens composed of the target virus and contaminants while the negative specimen contained only the contaminants. In such cases, the positive unlabeled classi cation (PUC) method 21 is used to remove the waveform of contaminants 20 . The decision on positive/negative is made by a method of assembling the accuracy of the obtained waveform.…”