Quantifying accuracy and heterogeneity in single-molecule super-resolution microscopy.

Area of Science:

• Biophysics
• Microscopy
• Computational Biology

Background:

• Single-molecule localization microscopy (SMLM) is crucial for nanoscale imaging.
• Current SMLM benchmarking methods lack the ability to quantify nanoscale accuracy on arbitrary datasets.
• There is a need for robust methods to assess localization confidence without ground-truth samples.

Purpose of the Study:

• To develop a computational method for robustly measuring the nanoscale accuracy and confidence of individual localizations in SMLM data.
• To validate the method's applicability across different SMLM datasets, imaging systems, and reconstruction algorithms.
• To demonstrate the utility of localization confidence in improving SMLM data analysis and interpretation.

Main Methods:

• Development of Wasserstein-induced flux (WIF), a computational technique measuring localization stability under system-informed perturbations.
• Application of WIF to experimental 2D and 3D SMLM data of microtubules and amyloid fibrils.
• Evaluation of WIF's performance in assessing reconstruction algorithm accuracy and enhancing structural resolution.

Main Results:

• WIF successfully measures the accuracy of various SMLM reconstruction algorithms on experimental data.
• The method provides robust localization confidence without requiring ground-truth sample information.
• WIF confidences enable evaluation of model-data mismatch and reveal hidden molecular heterogeneities.

Conclusions:

• Wasserstein-induced flux (WIF) offers a broadly applicable computational solution for quantifying SMLM data accuracy and confidence.
• WIF can improve the reliability and resolution of nanoscale imaging reconstructions.
• This methodology has the potential to advance SMLM data analysis across diverse biological applications.