A framework for evaluating the performance of SMLM cluster analysis algorithms

Nieves, Daniel J.; Ja, Pike; Griffié, Juliette; Sage, Daniel; Ea, Cohen; Sibarita, Jean‐Baptiste; Heilemann, Mike; Dm, Owen

doi:10.1101/2021.06.19.449098

Cited by 4 publications

(2 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Various clustering algorithms have been developed to quantify the degree or variability of clustering of a protein of interest under various conditions. Examples include local clustering algorithms that define boundaries of dense localizations ( Ester et al, 1996 ; Perry, 2004 ; Owen et al, 2010 ; Pageon et al, 2016a ; Griffié et al, 2016 ; Levet et al, 2019 ; Khater et al, 2020 ; Nino et al, 2020 ; Simoncelli et al, 2020 ; Williamson et al, 2020 ; Marenda et al, 2021 ; Nieves et al, 2021 ) or bulk metrics based on the radial distribution or pair-correlation function that quantify the density of localization pairs as a function of their distance to each other ( Ripley, 1979 ; Kiskowski et al, 2009 ; Sengupta et al, 2011 ; Veatch et al, 2012 ; Stone and Veatch, 2015 ). Importantly, these analysis methods can be expanded to two-color SMLM data to quantify the colocalization and structural relation of two proteins.…”

Section: Introductionmentioning

confidence: 99%

Efficient Cross-Correlation Filtering of One- and Two-Color Single Molecule Localization Microscopy Data

et al. 2021

View full text Add to dashboard Cite

Single molecule localization microscopy has become a prominent technique to quantitatively study biological processes below the optical diffraction limit. By fitting the intensity profile of single sparsely activated fluorophores, which are often attached to a specific biomolecule within a cell, the locations of all imaged fluorophores are obtained with ∼20 nm resolution in the form of a coordinate table. While rendered super-resolution images reveal structural features of intracellular structures below the optical diffraction limit, the ability to further analyze the molecular coordinates presents opportunities to gain additional quantitative insights into the spatial distribution of a biomolecule of interest. For instance, pair-correlation or radial distribution functions are employed as a measure of clustering, and cross-correlation analysis reveals the colocalization of two biomolecules in two-color SMLM data. Here, we present an efficient filtering method for SMLM data sets based on pair- or cross-correlation to isolate localizations that are clustered or appear in proximity to a second set of localizations in two-color SMLM data. In this way, clustered or colocalized localizations can be separately rendered and analyzed to compare other molecular properties to the remaining localizations, such as their oligomeric state or mobility in live cell experiments. Current matrix-based cross-correlation analyses of large data sets quickly reach the limitations of computer memory due to the space complexity of constructing the distance matrices. Our approach leverages k-dimensional trees to efficiently perform range searches, which dramatically reduces memory needs and the time for the analysis. We demonstrate the versatile applications of this method with simulated data sets as well as examples of two-color SMLM data. The provided MATLAB code and its description can be integrated into existing localization analysis packages and provides a useful resource to analyze SMLM data with new detail.

show abstract

Section: Introductionmentioning

confidence: 99%

Efficient Cross-Correlation Filtering of One- and Two-Color Single Molecule Localization Microscopy Data

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The importance of benchmarking clustering algorithms has recently been addressed by proposing a unifying framework for comparison 27 .…”

mentioning

confidence: 99%

Unbiased choice of global clustering parameters for single-molecule localization microscopy

Verzelli

Nold

Sun

et al. 2022

Sci Rep

Self Cite

View full text Add to dashboard Cite

Single-molecule localization microscopy resolves objects below the diffraction limit of light via sparse, stochastic detection of target molecules. Single molecules appear as clustered detection events after image reconstruction. However, identification of clusters of localizations is often complicated by the spatial proximity of target molecules and by background noise. Clustering results of existing algorithms often depend on user-generated training data or user-selected parameters, which can lead to unintentional clustering errors. Here we suggest an unbiased algorithm (FINDER) based on adaptive global parameter selection and demonstrate that the algorithm is robust to noise inclusion and target molecule density. We benchmarked FINDER against the most common density based clustering algorithms in test scenarios based on experimental datasets. We show that FINDER can keep the number of false positive inclusions low while also maintaining a low number of false negative detections in densely populated regions.

show abstract

Recent development of computational cluster analysis methods for single-molecule localization microscopy images

Hyun

Kim

2023

Computational and Structural Biotechnology Journal

View full text Add to dashboard Cite

A framework for evaluating the performance of SMLM cluster analysis algorithms

Cited by 4 publications

References 36 publications

Efficient Cross-Correlation Filtering of One- and Two-Color Single Molecule Localization Microscopy Data

Efficient Cross-Correlation Filtering of One- and Two-Color Single Molecule Localization Microscopy Data

Unbiased choice of global clustering parameters for single-molecule localization microscopy

Recent development of computational cluster analysis methods for single-molecule localization microscopy images

Contact Info

Product

Resources

About