U-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking data

Biophysical processes within living systems rely on encounters and interactions between molecules in complex environments such as cells. They are often described by anomalous diffusion transport. Recent advances in single-molecule microscopy and particle-tracking techniques have yielded an abundance...

Full description

Saved in:
Bibliographic Details
Main Authors: Solomon Asghar, Ran Ni, Giorgio Volpe
Format: Article
Language:English
Published: IOP Publishing 2025-01-01
Series:JPhys Photonics
Subjects:
Online Access:https://doi.org/10.1088/2515-7647/adf9aa
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849230231840227328
author Solomon Asghar
Ran Ni
Giorgio Volpe
author_facet Solomon Asghar
Ran Ni
Giorgio Volpe
author_sort Solomon Asghar
collection DOAJ
description Biophysical processes within living systems rely on encounters and interactions between molecules in complex environments such as cells. They are often described by anomalous diffusion transport. Recent advances in single-molecule microscopy and particle-tracking techniques have yielded an abundance of data in the form of videos and trajectories that contain critical information about these biologically significant processes. However, standard approaches for characterizing anomalous diffusion from these measurements often struggle in cases of practical interest, e.g. due to short, noisy trajectories. Fully exploiting this data therefore requires the development of advanced analysis methods—a core goal at the heart of the recent international Anomalous Diffusion (AnDi) Challenges. Here, we introduce a novel machine-learning framework, U-net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME), that applies a U-Net 3+ based neural network alongside Gaussian mixture models to enable highly accurate characterisation of single-particle tracking data. In the 2024 AnDi challenge, U-AnD-ME outperformed all other participating methods for the analysis of two-dimensional anomalous diffusion trajectories at both single-trajectory and ensemble levels. Using a large dataset inspired by the Challenge and experimental trajectories, we further characterize the performance of U-AnD-ME in segmenting trajectories and inferring anomalous diffusion properties.
format Article
id doaj-art-6db8b68bb542402f83a78aed30e681b5
institution Kabale University
issn 2515-7647
language English
publishDate 2025-01-01
publisher IOP Publishing
record_format Article
series JPhys Photonics
spelling doaj-art-6db8b68bb542402f83a78aed30e681b52025-08-21T06:57:46ZengIOP PublishingJPhys Photonics2515-76472025-01-017404500510.1088/2515-7647/adf9aaU-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking dataSolomon Asghar0Ran Ni1https://orcid.org/0000-0001-9478-0674Giorgio Volpe2https://orcid.org/0000-0001-9993-5348Department of Chemistry, University College London , 20 Gordon Street, WC1H 0AJ London, United Kingdom; School of Chemistry, Chemical Engineering and Biotechnology, Nanyang Technological University , Singapore 639798, SingaporeSchool of Chemistry, Chemical Engineering and Biotechnology, Nanyang Technological University , Singapore 639798, SingaporeDepartment of Chemistry, University College London , 20 Gordon Street, WC1H 0AJ London, United KingdomBiophysical processes within living systems rely on encounters and interactions between molecules in complex environments such as cells. They are often described by anomalous diffusion transport. Recent advances in single-molecule microscopy and particle-tracking techniques have yielded an abundance of data in the form of videos and trajectories that contain critical information about these biologically significant processes. However, standard approaches for characterizing anomalous diffusion from these measurements often struggle in cases of practical interest, e.g. due to short, noisy trajectories. Fully exploiting this data therefore requires the development of advanced analysis methods—a core goal at the heart of the recent international Anomalous Diffusion (AnDi) Challenges. Here, we introduce a novel machine-learning framework, U-net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME), that applies a U-Net 3+ based neural network alongside Gaussian mixture models to enable highly accurate characterisation of single-particle tracking data. In the 2024 AnDi challenge, U-AnD-ME outperformed all other participating methods for the analysis of two-dimensional anomalous diffusion trajectories at both single-trajectory and ensemble levels. Using a large dataset inspired by the Challenge and experimental trajectories, we further characterize the performance of U-AnD-ME in segmenting trajectories and inferring anomalous diffusion properties.https://doi.org/10.1088/2515-7647/adf9aaparticle trackingmachine learningmicroscopy data analysisanomalous diffusion
spellingShingle Solomon Asghar
Ran Ni
Giorgio Volpe
U-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking data
JPhys Photonics
particle tracking
machine learning
microscopy data analysis
anomalous diffusion
title U-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking data
title_full U-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking data
title_fullStr U-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking data
title_full_unstemmed U-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking data
title_short U-Net 3+ for anomalous diffusion analysis enhanced with mixture estimates (U-AnD-ME) in particle-tracking data
title_sort u net 3 for anomalous diffusion analysis enhanced with mixture estimates u and me in particle tracking data
topic particle tracking
machine learning
microscopy data analysis
anomalous diffusion
url https://doi.org/10.1088/2515-7647/adf9aa
work_keys_str_mv AT solomonasghar unet3foranomalousdiffusionanalysisenhancedwithmixtureestimatesuandmeinparticletrackingdata
AT ranni unet3foranomalousdiffusionanalysisenhancedwithmixtureestimatesuandmeinparticletrackingdata
AT giorgiovolpe unet3foranomalousdiffusionanalysisenhancedwithmixtureestimatesuandmeinparticletrackingdata