Nanopore read-out, that is the current signals measured across nanometer-sized openings in dielectric membranes or through natural protein channels, enables the detection, identification and sequencing of individual molecules. The detection can take place by analyzing the events of single biomolecules interacting with the pore. The accuracy in the detection of these single events is key for identification of physicochemical properties of analyte molecules. To this end, we further develop a very simple, fast, almost parameter-free, and adaptable cluster-based event detection (CBED) algorithm that clusters the nanopore signals prior to detecting nanopore events. The algorithm is validated against two other event detection schemes with respect to simplicity and efficiency. For this, nanopore data from four different experiments stemming from different laboratories that vary in the nanopore type, size, and analyte are considered. The comparison is made on the basis of the number of events detected, their quality, and the most important features extracted from nanopore events. Our results underline the higher efficiency and less noise of the CBED detected events for biological nanopore data and the need for an on-the-fly adaptivity of the baseline current for a class of solid-state nanopore data.

Nanopores enable single-molecule analysis by measuring current signals through nanoscale pores in either biological or solid-state membranes. Accurate detection of analyte fingerprints within the pore environment is essential for reading-out the analyte type. We develop a framework for robust and label-free detection of the molecular nanopore events using a graph representation of the measured signals. To this end, we build a graph-based two-stage workflow based on a convolutional and graph neural networks that first perform a fast screening of the nanopore events, followed by a deep validation of these. The learned model can thus efficiently and in an unsupervised manner select possible molecular signatures (the current blockades) in the full signal, denoise, validate, reconstruct these, and predict the morphology of unseen molecular events. We could show that the learned model can efficiently predict the correct event morphology for the same analyte within a 2.4-fold range of transmembrane voltage values not included in the training. The developed graph-based workflow is modular, generalizable, and provided that it is trained on a huge amount of different nanopore experiments has the potential to become a blueprint model for nanopore read-out. Such a read-out model would be able to identify subtle differences in molecules like proteins, as well as their conformational or folding states. The proposed framework is developed using experimental signals from DNA translocation through an aerolysin pore and demonstrates a unified approach linking unsupervised feature learning to raw-signal inference for single-molecule sensing.

Nanopore-based single-molecule sensing is a promising route to fast, low-cost disease diagnosis and protein sequencing: as an analyte such as a peptide or protein traverses a nanoscale pore, it modulates the ionic current, producing a resistive pulse whose signature is determined by the analyte's structure and its interactions with the pore. Translating these signatures into reliable molecular identities, however, is an open problem well suited for machine learning, as the signals are noisy, suffer from variations due to experimental conditions, and are difficult to featurize, which has so far limited classification accuracy. Here we translate the peptide identification problem into an image-classification task by transforming each resistive pulse into a scaleogram via the continuous wavelet transform, a representation that jointly encodes amplitude, frequency, and time in a form well suited for deep convolutional models. On a dataset of 42 peptides, recorded as six separate peptide ladders, this approach reaches a macro-averaged classification accuracy of 82 % on held-out events, an improvement of 8.6 percentage points over the descriptor-based approach previously reported for the same dataset. We further show that the trained models tolerate substantial compression, retaining their accuracy with half of their weights set to zero and under 8-bit quantization, a prerequisite for deploying trained classifiers on embedded sensing hardware. Our results demonstrate how physically motivated signal representations can make complex single-molecule data tractable for modern learning algorithms, a step on the path towards point-of-care peptide and protein diagnostics.

Peptide classification using nanopore-based devices promises to be a breakthrough method in basic research, diagnostics, and analytics. However, the measured blockage currents suffer from a low signal-to-noise ratio and a high information density that has hitherto not been fully deciphered. Some simple machine learning approaches using average current blockade depths and dwell-times have been investigated to improve this situation. In this work, a comprehensive statistical analysis of nanopore current signals is performed and demonstrated to be sufficient for classifying up to 42 peptides with over 70% accuracy. Two sets of features, the statistical moments and the catch22 set, are compared both in their representations and after training small classifier neural networks. We demonstrate that complex features of the events, captured in both the catch22 set and the central moments, are key to classifying peptides with otherwise similar mean currents. These results highlight the efficacy of purely statistical analysis of nanopore data and suggest a path forward for more sophisticated classification techniques.

The pore formed by the bacterial toxin aerolysin is valuable as a sensor of both natural and synthetic polymers through its ability to trap molecules in its interior for prolonged times, allowing precise measurements of the degree of conductance block. A prominent feature in these studies was that trapping of polymers occurred only from the cis-compartment from which the pore had inserted into the membrane. As some of these analytes are clearly transported through the pore, they appear able to pass the trans opening in one direction (cis-trans), but not the other (trans-cis).

In-vivo, transmembrane proteins are embedded in a complex matrix comprised of many different types of lipids, lipo-polysaccharides, small molecules, and other moieties. In contrast, many in-situ studies of ion channel structure and function are often performed with membrane mimics (in the liquid-crystalline state) made solely from homogeneous lipids, initially suspended in an organic solvent. To determine the effects of lipid type on the ionic conductance and geometry of a nanopore used extensively for biosensing application development, we measured the current-voltage relationship of the ion channel obtained from pre-formed heptamers of Staphylococcus aureus alpha-hemolysin in membranes comprised by either of two different phospholipids.

Protein characterization using nanopore-based devices promises to be a breakthrough method in basic research, diagnostics, and analytics. Current research includes the use of machine learning to achieve this task. In this work, a comprehensive statistical analysis of nanopore current signals is performed and demonstrated to be sufficient for classifying up to 42 peptides with 70 catch22 set, are compared both in their representations and after training small classifier neural networks. We demonstrate that complex features of the events, captured in both the catch22 set and the central moments, are key in classifying peptides with otherwise similar mean currents. These results highlight the efficacy of purely statistical analysis of nanopore data and suggest a path forward for more sophisticated classification techniques.

Citrullination, a particularly subtle (+1 Da) post-translational modification (PTM), has recently gained significant interest in the field of biomedicine because of its suspected involvement in human diseases like rheumatoid arthritis and tumours. In spite of minimal change in residue size introduction of citrulline in histone proteins is known to drastically change the structure and function. Hence, sensitive techniques are in demand for the detection of citrullination. In the light of recent demonstrations of the high sensitivity of the aerolysin pore to peptide volume (Piguet et al.

While DNA-sequencing by nanopores is established, there is no similar method available for proteins and peptides. First attempts towards peptide sequencing by nanopores have recently been undertaken by the use of a DNA-peptide hybrid threaded through a nanopore by a helicase DNA processive enzyme. However, this approach faces several intrinsic restrictions that hamper general application. Here, we present an alternative by peptide differentiation using a wt-Aerolysin pore in the trapping regime.

It has recently been shown that ionic current modulation by analytes visiting the pore formed by wild-type or mutant aerolysin (AeL) toxin is sensitive not only to molecular volume but also to molecular shape (1). One prerequisite of the remarkable sensitivity of this “whole molecule sensing” regime is the prolonged (1-100 ms) trapping of single molecules in the beta-barrel of the pore. At the same time, it is well known that macromolecular analytes are unable to enter this portion of the AeL-pore directly from the trans-side (2) but must enter from cis passinga constriction. We use high-bandwidth (0-100 kHz) recording of blockades induced by short polynucleotides (dA3,dA4,dA5) and oligoarginine peptides (R3,R5,R7) in conjunction with mutations of charged residues defining the trans-ward boundary of the constriction (D222C, R220S) to probe its role in the sensing mechanism. Both mutations induced rectification (D222C cis-ward; R220S trans-ward) and significantly affected the probability of transient deep blocks indicating analyte returns to the constriction after the first visit to the trap, with D222C potentiating and R220S reducing it for nuleotides. However, the initial deep block indicative of constriction passage remained in R220S. In the case of peptides, however, D222C left deep blocks largely unaffected while R220S nearly abolished returns and prolonged total blockage durations. These and other results indicate that the residues at the border between the constriction and the trap do not merely act by their charges but exert important steric influences and that further modifications of these residues may tailor the pore to various analytes. (1) Ouldali et al. 2020. Nat. Biotech. 38:176-181; Ensslen et al. 2022. JACS 144:16060-16068. (2) Baaken et al. 2015. ACS Nano. 9:6443-6449. (3) Iacovache et al. 2016. Nat Comms. 7:12062-8.

The standard model of pore formation was introduced more than fifty years ago, and it has been since, despite some refinements, the cornerstone for interpreting experiments related to pores in membranes. A central prediction of the model concerning pore opening under an electric field is that the activation barrier for pore formation is lowered proportionally to the square of the electric potential. However, this has only been scarcely and inconclusively confronted to experiments. In this paper, we study the electropermeability of model lipid membranes composed of 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine (POPC) containing different fractions of POPC-OOH, the hydroperoxidized form of POPC, in the range 0 to 100 mol %. By measuring ion currents across a 50-μm-diameter black lipid membrane (BLM) with picoampere and millisecond resolution, we detect hydroperoxidation-induced changes to the intrinsic bilayer electropermeability and to the probability of opening angstrom-size or larger pores. Our results over the full range of lipid compositions show that the energy barrier to pore formation is lowered linearly by the absolute value of the electric field, in contradiction with the predictions of the standard model.

Amino acid chirality may have significant roles in diseases such as Alzheimer's and certain forms of cancer. Today, LC-MS/MS and circular dichroism are used to distinguish L and D forms. However, no single molecule techniques are available to analyze chiral variability of peptides. Here, nanopores formed by the bacterial toxin aerolysin (AeL) were used to differentiate the three octaarginines RRRR-RRRR (R8), rrrr-rrrr (r8) and rrrr-RRRR (r4-R4), where R: L-R and r: D-R. Trans negative potentials were applied to pores reconstituted in DPhPC-bilayers in 4 M KCl. Using wild-type (wt) AeL, we found distinctive behaviours of the relative residual conductance of the peptide-blocked state (I/Io) with an increase in negative bias (−30 to −70 mV): I/Io decreased for R8 (0.357±0.002 to 0.352±0.002), remained constant for r8 (0.360±0.002) and increased for r4-R4 (0.353±0.001 to 0.359±0.001). As a consequence, the sequence of I/Io values was r8>R8>r4-R4 at −30 mV with R8 and r4-R4 trading places at about −45 mV. In contrast, with the R220S mutant of AeL, I/Io for R8 and r8 decreased in parallel from −30 to −70 mV (0.332±0.002 to 0.322±0.002 for R8 and 0.340±0.004 to 0.333±0.002 for r8) but remained constant for r4-R4 (0.332±0.002). Characteristic durations of blocked states were more than 2-fold enhanced with R220S vs. wt-AeL, were similar for R8 and r8 but, surprisingly, about 2.5-fold longer for r4-R4 in both pores, with maxima occurring at −30 mV (R220S) and −40 mV (wt). Ionic current through the aerolysin pore thus appears to be differentially modulated by L- and D-residues of peptides. Differential voltage-dependencies of I/Io and prolonged dwell time of the heterochiral peptide might be clues to a mechanism involving peptide conformation and warrant further experimental and theoretical study.

Fluorescence microscopy and, in particular, single molecule optical spectroscopy is of great potential value in characterizing the structural dynamics of membranes and membrane proteins. A particular challenge is to combine such high-resolution optical measurements with high-resolution voltage clamp electrical recordings that can provide direct information on single ion channel gating a block by drugs and analytes. Here, we report on the use of a novel chip-based, 2 × 2 arrangement of microelectrode cavities with borosilicate glass optical windows which facilitates optical access on an inverted microscope with water or oil-immersion objectives of high numerical aperture to horizontal free-standing lipid membranes, while controlling membrane voltage and recording currents using individual micropatterned, ring-shaped Ag/AgCl-electrodes to perform time-resolved single photon counting on free-standing membranes spanning sub-nanoliter cavities. Single channel activity induced by the fluorescently labelled pore-forming peptide ceratotoxin A was simultaneously acquired. This device allows for rapid formation of four membranes that are simultaneously monitored electrically using a four-channel amplifier and can be sequentially optically addressed, greatly reducing the time needed for successful experimentation. During our experiments, we noted autofluorescence of the borosilicate to be a limiting factor for wide-field low-intensity fluorescence recording, but this can likely be circumvented by using quartz glass instead. In summary, the novel device increases the likelihood of realizing the long standing ambition to correlate structural and functional dynamics of single membrane proteins on the single molecule level.

Posttranslational modifications (PTMs) are crucial for cellular function of proteins but pose analytical problems, especially when distingushing chemically identical PTMs placed at different locations within the same protein. Current methods, such as liquid-chromatography-tandem mass spectrometry (LC-MS/MS), are technically tantamount to de novo protein sequencing. Nanopore sensing is an emerging potential alternative, but sequencing proteins using the methodology of nanopore DNA sequencing still faces fundamental problems.

Protein sequencing has been an important goal in proteomics and diagnostics for decades. Current techniques rely on cost intensive large-scale equipment, such as HPLC coupled to mass-spectrometry and subsequent in-silico big-data-analysis. These approaches are also time consuming and error prone. Here, we present a novel approach for peptide sequence recognition as a major step towards sequencing in a derivatization-free single molecule experiment using the wt-aerolysin (wt-AeL) nanopore. We follow a bottom-up peptide ladder strategy, which we validate using six different peptide ladder-like sample pools. Each pool is based on a hetero-deca-peptide consisting of a scrambled heptameric sequence of 5 amino acids (S,R,K,Y,A) followed by a C-terminal tri-arginine carrier and -in addition to the full length peptide- contains six fragments shortened by one to six amino acids starting at the N-terminus. We show that using this strategy, sequences can be identified on the basis of resistive pulse analysis using the wt-AeL pore. We also provide evidence that the influence of a single N-terminal amino-acid (aa) removal/addition on resistive pulse depth is sensitive to both its species and that of its nearest neighbor, thus opening an avenue towards peptide sequencing by nanopore.

Optical techniques, such as fluorescence microscopy, are of great value in characterizing the structural dynamics of membranes and membrane proteins. A particular challenge is to combine high-resolution optical measurements with high-resolution voltage clamp electrical recordings providing direct information on e.g. single ion channel gating and/or membrane capacitance. Here, we report on a novel chip-based array device which facilitates optical access with water or oil-immersion objectives of high numerical aperture to horizontal free-standing lipid membranes while controlling membrane voltage and recording currents using individual micropatterned Ag/AgCl-electrodes. Wide-field and confocal imaging, as well as time-resolved single photon counting on free-standing membranes spanning sub-nanoliter cavities are demonstrated while electrical signals, including single channel activity, are simultaneously acquired. This optically addressable microelectrode cavity array will allow combined electrical-optical studies of membranes and membrane proteins to be performed as a routine experiment.