Abstract:Quantifying the number of molecules from fluorescence microscopy measurements is an important topic in cell biology and medical research. In this work, we present a consecutive algorithm for super-resolution (STED) scanning microscopy that provides molecule counts in automatically generated image segments and offers statistical guarantees in form of asymptotic confidence intervals. To this end, we first apply a multiscale scanning procedure on STED microscopy measurements of the sample to obtain a system of significant regions, each of which contains at least one molecule with prescribed uniform probability. This system of regions will typically be highly redundant and consists of rectangular building blocks. To choose an informative but non-redundant subset of more naturally shaped regions, we hybridize our system with the result of a generic segmentation algorithm. The diameter of the segments can be of the order of the resolution of the microscope. Using multiple photon coincidence measurements of the same sample in confocal mode, we are then able to estimate the brightness and number of the molecules and give uniform confidence intervals on the molecule counts for each previously constructed segment. In other words, we establish a so-called molecular map with uniform error control. The performance of the algorithm is investigated on simulated and real data.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the quantification of the number of molecules in a sample under super - resolution microscopy. Specifically, the authors propose a sequential algorithm for super - resolution (STED) scanning microscopy, which can provide molecule counts in automatically generated image segments and provide statistical guarantees in the form of asymptotic confidence intervals. This helps to more precisely understand the molecule distribution and quantity in samples in cell biology and medical research. ### Main contributions of the paper: 1. **Development of the statistical model**: - The authors develop a detailed statistical model for the contribution of any high - order photon coincidence from the observational data obtained by anti - condenser microscopes. - Unlike traditional global estimation methods, this model constructs a so - called "molecule graph", which contains multiple independent segments with uniform error control, and each segment has a confidence interval for the number of molecules. 2. **Construction of the molecule graph**: - The molecule graph is constructed by a three - step method: - **Segmentation**: First, the image space is segmented so that each segment contains at least one molecule and has an appropriate shape. - **Estimation**: Based on the segmentation results, the number of molecules is estimated segment by segment. - **Confidence interval**: Based on the central limit theorem, a confidence interval for the number of molecules in each segment is constructed. 3. **Technical details**: - A multi - scale scanning process is used to process the samples measured by STED microscopes to obtain a set of significant regions, each of which contains at least one molecule. - By mixing the results of general - purpose segmentation algorithms, a subset of natural - shape regions that are information - rich and non - redundant is selected. - Multiple photon coincidence measurements are used to estimate the brightness and number of molecules and provide a unified confidence interval for the molecule count for each previously constructed segment. ### Formula explanation: - **Photon detection probability**: \[ p_j(x) = p_j\cdot h(x - x_{\psi(j)}) \] where \( p_j \) is the brightness of the \( j \) - th marker, \( h \) is the point spread function (PSF) of the microscope, and \( x_{\psi(j)} \) is the position of the \( j \) - th marker. - **Convolution model**: \[ g(x)=\sum_{j = 1}^N p_j(x)=\int_{[0,1]^2} f_N(y)f_p(y)h(x - y)\,dy \] where \( g(x) \) represents the probability of detecting at least one photon when scanning at position \( x \), and \( f_N \) and \( f_p \) represent the spatial number and brightness of the markers respectively. - **Polynomial distribution**: \[ Y_i=(Y_0^i,\ldots,Y_{m_d}^i)\sim \text{M}(t,E_0(x_i),\ldots,E_{m_d}(x_i)),\quad i\in [n]\times [n] \] where \( Y_k^i \) represents the number of events in which \( k \) photons are detected at position \( x_i \), \( t \) is the total number of experiments per pixel, and \( E_k(x_i) \) is the probability of observing \( k \) photons at position \( x_i \). Through these methods and techniques, the paper aims to provide a reliable and statistically - guaranteed molecule counting method, thereby achieving more accurate quantitative analysis under super - resolution microscopy.

Towards quantitative super-resolution microscopy: Molecular maps with statistical guarantees

Toward quantitative super-resolution microscopy: molecular maps with statistical guarantees

Sub-Diffraction Imaging with Confocal Fluorescence Microscopy by Stochastic Photobleaching

Simultaneous super-resolution estimation of single-molecule position and orientation with minimal photon fluxes

3D resoution enhancement in saturated competition microscopy

Stochastic Approach to the Molecular Counting Problem in Superresolution Microscopy.

A correlation analysis framework for localization-based super-resolution microscopy

Advances in Sub-20-nm Fluorescence Super-Resolution Microscopy(Invited)

Statistical Molecule Counting in Super-Resolution Fluorescence Microscopy: Towards Quantitative Nanoscopy

Correlation Analysis Framework for Localization-Based Superresolution Microscopy.

Microtubule tracking from stochastic optical reconstruction microscopy images

Super-Resolution Microscopy Based on the Inherent Fluctuations of Dye Molecules

High-density volumetric super-resolution microscopy

Video-Rate Super Resolution Microscopy in Living Cells

Quantitative Single-Molecule Imaging by Confocal Laser Scanning Microscopy

Photons in - numbers out: Perspectives in quantitative fluorescence microscopy for in situ protein counting

Super-resolution capacity of variance-based stochastic fluorescence microscopy

Super-resolution fluorescence microscopy by line-scanning with an unmodified two-photon microscope

Three-Dimensional Super-Resolution Imaging by Stochastic Optical Reconstruction Microscopy

Pixelation with concentration-encoded effective photons for molecular optical sectioning microscopy

Analysis of Super-resolution Single Molecule Localization Microscopy Data: a tutorial