Machine Learning to Improve the Sensing of Biomolecules by Conical Track-Etched Nanopore

Abstract

Single nanopore is a powerful platform to detect, discriminate and identify biomacromolecules. Among the different devices, the conical nanopores obtained by the track-etched technique on a polymer film are stable and easy to functionalize. However, these advantages are hampered by their high aspect ratio that avoids the discrimination of similar samples. Using machine learning, we demonstrate an improved resolution so that it can identify short single- and double-stranded DNA (10- and 40-mers). We have characterized each current blockade event by the relative intensity, dwell time, surface area and both the right and left slope. We show an overlap of the relative current blockade amplitudes and dwell time distributions that prevents their identification. We define the different parameters that characterize the events as features and the type of DNA sample as the target. By applying support-vector machines to discriminate each sample, we show accuracy between 50% and 72% by using two features that distinctly classify the data points. Finally, we achieved an increased accuracy (up to 82%) when five features were implemented.

A Guide to Signal Processing Algorithms for Nanopore Sensors

“…Afterward, classification is carried out using SVMs with 96% accuracy to discriminate two highly similar analytes. Along the same lines, 73 each current blockade event can be characterized by the relative intensity, duration, surface area, and both the right and left slope of the pulses. The different parameters characterizing the events are defined as features and the type of DNA sample as the target.…”

Resistive‐pulse Sensing of DNA with a Polymeric Nanopore Sensor and Characterization of DNA Translocation

“…Besides having adjustable pore geometry and size; track‐etched polymer membranes are also preferred for the ease of surface modification for enhanced and selective sensing [6c,14] . Also DNA sensing with track‐etched nanopore was integrated with machine learning studies for differentiation of molecules and data analysis [15] …”

The NEOtrap – en route with a new single-molecule technique

“…A library of fingerprints can be collected (Figure 3B) and used as a training set for pattern recognition algorithms to facilitate the label-free identification of protein mixtures. Such machine-learningenhanced sensing approaches (Arima et al, 2018;Cao et al, 2020;Meyer et al, 2020) may provide a basis for single-cell proteomics in the future. In all this, the label-free aspect is a key advantage of the NEOtrap as that provides the option of direct measurements of scarce and nonpurified biological samples, such as lysate, blood, sweat, saliva (Galenkamp et al, 2018;Sze et al, 2017) without additional preprocessing.…”