High Accuracy Protein Identification: Fusion of solid-state nanopore sensing and machine learning

TL;DR

This paper demonstrates high-accuracy protein identification by combining solid-state nanopore sensing at high bandwidths with machine learning and advanced data analysis, enabling differentiation of similarly sized proteins.

Contribution

It introduces a novel integrated approach using high-bandwidth nanopore sensing and machine learning for precise protein identification, surpassing previous limitations.

Findings

Achieved up to 88.7% accuracy in protein identification

Used the highest bandwidth (10 MHz) for nanopore protein sensing

Improved specificity to 96.4% with advanced clustering and analysis

Abstract

Proteins are arguably the most important class of biomarkers for health diagnostic purposes. Label-free solid-state nanopore sensing is a versatile technique for sensing and analysing biomolecules such as proteins at single-molecule level. While molecular-level information on size, shape, and charge of proteins can be assessed by nanopores, the identification of proteins with comparable sizes remains a challenge. Here, we present methods that combine solid-state nanopore sensing with machine learning to address this challenge. We assess the translocations of four similarly sized proteins using amplifiers with bandwidths (BWs) of 100 kHz (sampling rate=200 ksps) and 10 MHz (sampling rate=40 Msps), the highest bandwidth reported for protein sensing, using nanopores fabricated in <10 nm thick silicon nitride membranes. F-values of up to 65.9% and 83.2% (without clustering of the protein signals) were achieved with 100 kHz and 10 MHz BW instruments, respectively, for identification of the four proteins. The accuracy of protein identification was significantly improved by grouping the signals into several clusters depending on the event features, resulting in F-value and specificity reaching as high as 88.7% and 96.4%, respectively, for combinations of four proteins. The combined improvement in sensor signals through the use of high bandwidth instruments, advanced clustering, machine learning, and other advanced data analysis methods allows identification of proteins with high accuracy.