Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis

Autores
Saavedra, Lucas A.; Barrantes, Francisco J.
Año de publicación
2025
Idioma
español castellano
Tipo de recurso
artículo
Estado
versión publicada
Descripción
Fil: Saavedra, Lucas A. Pontificia Universidad Católica Argentina. Facultad de Ciencias Médicas; Argentina
Fil: Barrantes, Francisco J. Pontificia Universidad Católica Argentina. Facultad de Ciencias Médicas; Argentina
The application of deep learning (DL) approaches to the study of particle diffusion is becoming increasingly widespread, achieving highly accurate predictive performance that is often challenging to attain with traditional analytical tools. DL has made successful incursions in the field of single-particle tracking of biological molecules in 2- and 3-dimensional spaces. In this work, we introduce a WaveNet-based deep neural network (WadNet)- and convolution-based DL network coined Wad temporal convolutional networks (WadTCN) and its transformer-based variant to decode single-value or pointwise prediction of dynamic properties of bidimensional trajectories. We trained WadTCN to be tested in all 1st and 2nd Andi Challenge tasks with bidimensional trajectories, examined how WadTCN performed when learning is transferred from one network to another to accelerate training, and evaluated its capacity to segment trajectories with a data-driven approach, at variance with other works that rely on mean changes or parameter thresholding to determine change points. WadTCN was further compared to a pure temporal convolutional network and RANDI, a top-performing recurrent neural network in the 1st Andi Challenge. We also assessed the ability of WadTCN to classify and segment single-particle trajectories of a membrane-associated protein, the nicotinic acetylcholine receptor, in the plane of a live cell plasma membrane. WadTCN outperformed similar analytical tools in predicting anomalous diffusion exponents and demonstrated notable robustness in more challenging tasks like the segmentation of trajectories into two states and prediction of the diffusion coefficient of diffusing molecules. Furthermore, most layers of WadTCN perform as general feature extractors, suggesting its applicability in other tasks requiring training acceleration.
Fuente
JPhys Photonics. 7, 2025.
Materia
APRENDIZAJE PROFUNDO
DIFUSION DE PARTICULAS
MODELOS DE SIMULACION
Nivel de accesibilidad
acceso abierto
Condiciones de uso
https://creativecommons.org/licenses/by-nc-sa/4.0/
Repositorio
Repositorio Institucional (UCA)
Institución
Pontificia Universidad Católica Argentina
OAI Identificador
oai:ucacris:123456789/20082

id RIUCA_9cb5eb624a522215bf6ae8dae434e8c0
oai_identifier_str oai:ucacris:123456789/20082
network_acronym_str RIUCA
repository_id_str 2585
network_name_str Repositorio Institucional (UCA)
spelling Temporal convolutional networks work as general feature extractors for single-particle diffusion analysisSaavedra, Lucas A.Barrantes, Francisco J.APRENDIZAJE PROFUNDODIFUSION DE PARTICULASMODELOS DE SIMULACIONFil: Saavedra, Lucas A. Pontificia Universidad Católica Argentina. Facultad de Ciencias Médicas; ArgentinaFil: Barrantes, Francisco J. Pontificia Universidad Católica Argentina. Facultad de Ciencias Médicas; ArgentinaThe application of deep learning (DL) approaches to the study of particle diffusion is becoming increasingly widespread, achieving highly accurate predictive performance that is often challenging to attain with traditional analytical tools. DL has made successful incursions in the field of single-particle tracking of biological molecules in 2- and 3-dimensional spaces. In this work, we introduce a WaveNet-based deep neural network (WadNet)- and convolution-based DL network coined Wad temporal convolutional networks (WadTCN) and its transformer-based variant to decode single-value or pointwise prediction of dynamic properties of bidimensional trajectories. We trained WadTCN to be tested in all 1st and 2nd Andi Challenge tasks with bidimensional trajectories, examined how WadTCN performed when learning is transferred from one network to another to accelerate training, and evaluated its capacity to segment trajectories with a data-driven approach, at variance with other works that rely on mean changes or parameter thresholding to determine change points. WadTCN was further compared to a pure temporal convolutional network and RANDI, a top-performing recurrent neural network in the 1st Andi Challenge. We also assessed the ability of WadTCN to classify and segment single-particle trajectories of a membrane-associated protein, the nicotinic acetylcholine receptor, in the plane of a live cell plasma membrane. WadTCN outperformed similar analytical tools in predicting anomalous diffusion exponents and demonstrated notable robustness in more challenging tasks like the segmentation of trajectories into two states and prediction of the diffusion coefficient of diffusing molecules. Furthermore, most layers of WadTCN perform as general feature extractors, suggesting its applicability in other tasks requiring training acceleration.IOP Publishing2025info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionhttp://purl.org/coar/resource_type/c_6501info:ar-repo/semantics/articuloapplication/pdfhttps://repositorio.uca.edu.ar/handle/123456789/200822515-764710.1088/2515-7647/adbec8JPhys Photonics. 7, 2025.reponame:Repositorio Institucional (UCA)instname:Pontificia Universidad Católica Argentinaspainfo:eu-repo/semantics/openAccesshttps://creativecommons.org/licenses/by-nc-sa/4.0/2026-11-05T12:34:03Zoai:ucacris:123456789/20082instacron:UCAInstitucionalhttps://repositorio.uca.edu.ar/Universidad privadaNo correspondehttps://repositorio.uca.edu.ar/oaiclaudia_fernandez@uca.edu.arArgentinaNo correspondeNo correspondeNo correspondeopendoar:25852026-11-05 12:34:04.293Repositorio Institucional (UCA) - Pontificia Universidad Católica Argentinafalse
dc.title.none.fl_str_mv Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
title Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
spellingShingle Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
Saavedra, Lucas A.
APRENDIZAJE PROFUNDO
DIFUSION DE PARTICULAS
MODELOS DE SIMULACION
title_short Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
title_full Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
title_fullStr Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
title_full_unstemmed Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
title_sort Temporal convolutional networks work as general feature extractors for single-particle diffusion analysis
dc.creator.none.fl_str_mv Saavedra, Lucas A.
Barrantes, Francisco J.
author Saavedra, Lucas A.
author_facet Saavedra, Lucas A.
Barrantes, Francisco J.
author_role author
author2 Barrantes, Francisco J.
author2_role author
dc.subject.none.fl_str_mv APRENDIZAJE PROFUNDO
DIFUSION DE PARTICULAS
MODELOS DE SIMULACION
topic APRENDIZAJE PROFUNDO
DIFUSION DE PARTICULAS
MODELOS DE SIMULACION
dc.description.none.fl_txt_mv Fil: Saavedra, Lucas A. Pontificia Universidad Católica Argentina. Facultad de Ciencias Médicas; Argentina
Fil: Barrantes, Francisco J. Pontificia Universidad Católica Argentina. Facultad de Ciencias Médicas; Argentina
The application of deep learning (DL) approaches to the study of particle diffusion is becoming increasingly widespread, achieving highly accurate predictive performance that is often challenging to attain with traditional analytical tools. DL has made successful incursions in the field of single-particle tracking of biological molecules in 2- and 3-dimensional spaces. In this work, we introduce a WaveNet-based deep neural network (WadNet)- and convolution-based DL network coined Wad temporal convolutional networks (WadTCN) and its transformer-based variant to decode single-value or pointwise prediction of dynamic properties of bidimensional trajectories. We trained WadTCN to be tested in all 1st and 2nd Andi Challenge tasks with bidimensional trajectories, examined how WadTCN performed when learning is transferred from one network to another to accelerate training, and evaluated its capacity to segment trajectories with a data-driven approach, at variance with other works that rely on mean changes or parameter thresholding to determine change points. WadTCN was further compared to a pure temporal convolutional network and RANDI, a top-performing recurrent neural network in the 1st Andi Challenge. We also assessed the ability of WadTCN to classify and segment single-particle trajectories of a membrane-associated protein, the nicotinic acetylcholine receptor, in the plane of a live cell plasma membrane. WadTCN outperformed similar analytical tools in predicting anomalous diffusion exponents and demonstrated notable robustness in more challenging tasks like the segmentation of trajectories into two states and prediction of the diffusion coefficient of diffusing molecules. Furthermore, most layers of WadTCN perform as general feature extractors, suggesting its applicability in other tasks requiring training acceleration.
description Fil: Saavedra, Lucas A. Pontificia Universidad Católica Argentina. Facultad de Ciencias Médicas; Argentina
publishDate 2025
dc.date.none.fl_str_mv 2025
dc.type.none.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
http://purl.org/coar/resource_type/c_6501
info:ar-repo/semantics/articulo
format article
status_str publishedVersion
dc.identifier.none.fl_str_mv https://repositorio.uca.edu.ar/handle/123456789/20082
2515-7647
10.1088/2515-7647/adbec8
url https://repositorio.uca.edu.ar/handle/123456789/20082
identifier_str_mv 2515-7647
10.1088/2515-7647/adbec8
dc.language.none.fl_str_mv spa
language spa
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
https://creativecommons.org/licenses/by-nc-sa/4.0/
eu_rights_str_mv openAccess
rights_invalid_str_mv https://creativecommons.org/licenses/by-nc-sa/4.0/
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv IOP Publishing
publisher.none.fl_str_mv IOP Publishing
dc.source.none.fl_str_mv JPhys Photonics. 7, 2025.
reponame:Repositorio Institucional (UCA)
instname:Pontificia Universidad Católica Argentina
reponame_str Repositorio Institucional (UCA)
collection Repositorio Institucional (UCA)
instname_str Pontificia Universidad Católica Argentina
repository.name.fl_str_mv Repositorio Institucional (UCA) - Pontificia Universidad Católica Argentina
repository.mail.fl_str_mv claudia_fernandez@uca.edu.ar
_version_ 1881034335737348096
score 12.487082