Graph-based learning of nanopore translocation event signals

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 and sequence. Towards these aims, we develop a framework for robust, semi-supervised detection of molecular nanopore events on the basis of a graph representation of the measured signals. To this end, using experimental signals from DNA translocation through an aerolysin pore, we build a graph-based two-stage workflow employing convolutional and graph neural networks that first perform fast screening followed by deep validation of these nanopore events. The learned model is able to select possible molecular signatures (the current blockades) in the full signal, denoise, validate, and reconstruct them, and predict the morphology of unseen molecular events within the range of the available data. The trained 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 partially supervised workflow, linking unsupervised graph representation-learning to raw-signal inference is - at its current state - a computational proof of concept that proposes a direction for the development of modular and generalizable event detection workflows, provided that they are trained on a huge amount of different nanopore experiments.

Authors

Institutions

Publication Details

Journal
Nanoscale
Published
2026-11-16
DOI
https://doi.org/10.1039/d6nr01686c
Primary Topic
Nanopore and Nanochannel Transport Studies
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Graph-based learning of nanopore translocation event signals

Jan Claas Behrends, Tobias Ensslen, Mayukh Kansari, Maria Fyta
1 links
Nanoscale
Nanopore and Nanochannel Transport Studies
article

Graph-based learning of nanopore translocation event signals

Jan Claas Behrends, Tobias Ensslen, Mayukh Kansari, Maria Fyta
article en

Abstract

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 and sequence. Towards these aims, we develop a framework for robust, semi-supervised detection of molecular nanopore events on the basis of a graph representation of the measured signals. To this end, using experimental signals from DNA translocation through an aerolysin pore, we build a graph-based two-stage workflow employing convolutional and graph neural networks that first perform fast screening followed by deep validation of these nanopore events. The learned model is able to select possible molecular signatures (the current blockades) in the full signal, denoise, validate, and reconstruct them, and predict the morphology of unseen molecular events within the range of the available data. The trained 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 partially supervised workflow, linking unsupervised graph representation-learning to raw-signal inference is - at its current state - a computational proof of concept that proposes a direction for the development of modular and generalizable event detection workflows, provided that they are trained on a huge amount of different nanopore experiments.

Nanoscale
University of Freiburg (DE), University Medical Center Freiburg (DE), RWTH Aachen University (DE)
Deutsche Forschungsgemeinschaft, Bundesministerium für Bildung und Forschung
Nanopore and Nanochannel Transport Studies
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon