A framework for evaluating the performance of SMLM cluster analysis algorithms

This analysis compares the performance of seven algorithms for cluster analysis of single-molecule localization microscopy data. The results provide a framework for comparing these types of methods and point users to the best tools. Single-molecule localization microscopy (SMLM) generates data in the form of coordinates of localized fluorophores. Cluster analysis is an attractive route for extracting biologically meaningful information from such data and has been widely applied. Despite a range of cluster analysis algorithms, there exists no consensus framework for the evaluation of their performance. Here, we use a systematic approach based on two metrics to score the success of clustering algorithms in simulated conditions mimicking experimental data. We demonstrate the framework using seven diverse analysis algorithms: DBSCAN, ToMATo, KDE, FOCAL, CAML, ClusterViSu and SR-Tesseler. Given that the best performer depended on the underlying distribution of localizations, we demonstrate an analysis pipeline based on statistical similarity measures that enables the selection of the most appropriate algorithm, and the optimized analysis parameters for real SMLM data. We propose that these standard simulated conditions, metrics and analysis pipeline become the basis for future analysis algorithm development and evaluation.

Automated summary

This study compares the performance of seven algorithms for cluster analysis of single-molecule localization microscopy data. The results provide a framework for comparing these methods and guide users to the best tools. Cluster analysis is an effective method for extracting meaningful information from single-molecule localization microscopy data, but there is no consensus framework for evaluating the performance of different algorithms. This study proposes a systematic approach and evaluation metrics based on simulated conditions to score the success of clustering algorithms.