FOCAL3D: A 3-dimensional clustering package for single-molecule localization microscopy.

Area of Science:

• Biophysics
• Cell Biology
• Microscopy

Background:

• Single-molecule localization microscopy (SMLM) generates massive datasets for nanoscale biological research.
• Efficient quantification tools are needed to analyze complex SMLM data.
• Existing clustering algorithms like DBSCAN can be sensitive to parameter choices.

Purpose of the Study:

• To introduce FOCAL3D, a novel, fast, and accurate density-based algorithm for quantifying spatial clustering in large 3D SMLM datasets.
• To demonstrate FOCAL3D's objective parameter determination and robustness to noise and varying cluster sizes.
• To validate FOCAL3D's performance on simulated and biological data, including nuclear pore complex imaging.

Main Methods:

• Development of FOCAL3D, a linear-scaling, density-based clustering algorithm for 3D SMLM data.
• Validation using simulated datasets with varying noise levels and cluster characteristics.
• Application to 3D astigmatic dSTORM images of the nuclear pore complex (NPC).

Main Results:

• FOCAL3D demonstrates parametric insensitivity and objective parameter optimization, unlike DBSCAN.
• High accuracy, as indicated by F1 and Silhouette scores, even with significant noise and mixed cluster populations.
• Successful clustering of complex, heterogeneous 3D structures in biological SMLM data.

Conclusions:

• FOCAL3D provides an accurate, flexible, and efficient solution for analyzing large 3D SMLM datasets.
• The algorithm's robustness and objective parameter selection make it a valuable tool for nanoscale biological research.
• FOCAL3D is available as an open-source Python package.