Dataset Information

Deep-Channel uses deep neural networks to detect single-molecule events from patch-clamp data.

ABSTRACT: Single-molecule research techniques such as patch-clamp electrophysiology deliver unique biological insight by capturing the movement of individual proteins in real time, unobscured by whole-cell ensemble averaging. The critical first step in analysis is event detection, so called "idealisation", where noisy raw data are turned into discrete records of protein movement. To date there have been practical limitations in patch-clamp data idealisation; high quality idealisation is typically laborious and becomes infeasible and subjective with complex biological data containing many distinct native single-ion channel proteins gating simultaneously. Here, we show a deep learning model based on convolutional neural networks and long short-term memory architecture can automatically idealise complex single molecule activity more accurately and faster than traditional methods. There are no parameters to set; baseline, channel amplitude or numbers of channels for example. We believe this approach could revolutionise the unsupervised automatic detection of single-molecule transition events in the future.

SUBMITTER: Celik N

PROVIDER: S-EPMC6946689 | biostudies-literature | 2020 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep-Channel uses deep neural networks to detect single-molecule events from patch-clamp data.

Celik Numan N O'Brien Fiona F Brennan Sean S Rainbow Richard D RD Dart Caroline C Zheng Yalin Y Coenen Frans F Barrett-Jolley Richard R

Communications biology 20200107 1

Single-molecule research techniques such as patch-clamp electrophysiology deliver unique biological insight by capturing the movement of individual proteins in real time, unobscured by whole-cell ensemble averaging. The critical first step in analysis is event detection, so called "idealisation", where noisy raw data are turned into discrete records of protein movement. To date there have been practical limitations in patch-clamp data idealisation; high quality idealisation is typically laboriou ...[more]

PMID: 31925311

Similar Datasets

Project description:BackgroundThere is only one established drug binding site on sodium channels. However, drug binding of sodium channels shows extreme promiscuity: ∼25% of investigated drugs have been found to potently inhibit sodium channels. The structural diversity of these molecules suggests that they may not share the binding site, and/or the mode of action. Our goal was to attempt classification of sodium channel inhibitors by measuring multiple properties of inhibition in electrophysiology experiments. We also aimed to investigate if different properties of inhibition correlate with specific chemical properties of the compounds.Methodology/principal findingsA comparative electrophysiological study of 35 compounds, including classic sodium channel inhibitors (anticonvulsants, antiarrhythmics and local anesthetics), as well as antidepressants, antipsychotics and neuroprotective agents, was carried out using rNav1.2 expressing HEK-293 cells and the QPatch automatic patch-clamp instrument. In the multi-dimensional space defined by the eight properties of inhibition (resting and inactivated affinity, potency, reversibility, time constants of onset and offset, use-dependence and state-dependence), at least three distinct types of inhibition could be identified; these probably reflect distinct modes of action. The compounds were clustered similarly in the multi-dimensional space defined by relevant chemical properties, including measures of lipophilicity, aromaticity, molecular size, polarity and electric charge. Drugs of the same therapeutic indication typically belonged to the same type. We identified chemical properties, which were important in determining specific properties of inhibition. State-dependence correlated with lipophilicity, the ratio of the neutral form of molecules, and aromaticity: We noticed that the highly state dependent inhibitors had at least two aromatic rings, logP>4.0, and pKa<8.0.Conclusions/significanceThe correlations of inhibition properties both with chemical properties and therapeutic profiles would not have been evident through the sole determination of IC(50); therefore, recording multiple properties of inhibition may allow improved prediction of therapeutic usefulness.

Dataset Information

Deep-Channel uses deep neural networks to detect single-molecule events from patch-clamp data.

Publications

Deep-Channel uses deep neural networks to detect single-molecule events from patch-clamp data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets