Label-free identification of protein aggregates using deep learning.

Protein misfolding and aggregation play central roles in the pathogenesis of various neurodegenerative diseases (NDDs), including Huntington's disease, which is caused by a genetic mutation in exon 1 of the Huntingtin protein (Httex1). The fluorescent labels commonly used to visualize and monitor the dynamics of protein expression have been shown to alter the biophysical properties of proteins and the final ultrastructure, composition, and toxic properties of the formed aggregates. To overcome this limitation, we present a method for label-free identification of NDD-associated aggregates (LINA). Our approach utilizes deep learning to detect unlabeled and unaltered Httex1 aggregates in living cells from transmitted-light images, without the need for fluorescent labeling. Our models are robust across imaging conditions and on aggregates formed by different constructs of Httex1. LINA enables the dynamic identification of label-free aggregates and measurement of their dry mass and area changes during their growth process, offering high speed, specificity, and simplicity to analyze protein aggregation dynamics and obtain high-fidelity information.

Fluorescence Microscopy: a statistics-optics perspective.

Fundamental properties of light unavoidably impose features on images collected using fluorescence microscopes. Modeling these features is ever more important in quantitatively interpreting microscopy images collected at scales on par or smaller than light's wavelength. Here we review the optics responsible for generating fluorescent images, fluorophore properties, microscopy modalities leveraging properties of both light and fluorophores, in addition to the necessarily probabilistic modeling tools imposed by the stochastic nature of light and measurement.

Correlative 3D microscopy of single cells using super-resolution and scanning ion-conductance microscopy.

High-resolution live-cell imaging is necessary to study complex biological phenomena. Modern fluorescence microscopy methods are increasingly combined with complementary, label-free techniques to put the fluorescence information into the cellular context. The most common high-resolution imaging approaches used in combination with fluorescence imaging are electron microscopy and atomic-force microscopy (AFM), originally developed for solid-state material characterization. AFM routinely resolves atomic steps, however on soft biological samples, the forces between the tip and the sample deform the fragile membrane, thereby distorting the otherwise high axial resolution of the technique. Here we present scanning ion-conductance microscopy (SICM) as an alternative approach for topographical imaging of soft biological samples, preserving high axial resolution on cells. SICM is complemented with live-cell compatible super-resolution optical fluctuation imaging (SOFI). To demonstrate the capabilities of our method we show correlative 3D cellular maps with SOFI implementation in both 2D and 3D with self-blinking dyes for two-color high-order SOFI imaging. Finally, we employ correlative SICM/SOFI microscopy for visualizing actin dynamics in live COS-7 cells with subdiffraction-resolution.

Parameter-free rendering of single-molecule localization microscopy data for parameter-free resolution estimation.

Localization microscopy is a super-resolution imaging technique that relies on the spatial and temporal separation of blinking fluorescent emitters. These blinking events can be individually localized with a precision significantly smaller than the classical diffraction limit. This sub-diffraction localization precision is theoretically bounded by the number of photons emitted per molecule and by the sensor noise. These parameters can be estimated from the raw images. Alternatively, the resolution can be estimated from a rendered image of the localizations. Here, we show how the rendering of localization datasets can influence the resolution estimation based on decorrelation analysis. We demonstrate that a modified histogram rendering, termed bilinear histogram, circumvents the biases introduced by Gaussian or standard histogram rendering. We propose a parameter-free processing pipeline and show that the resolution estimation becomes a function of the localization density and the localization precision, on both simulated and state-of-the-art experimental datasets.

Smoothness correction for better SOFI imaging.

Sub-diffraction or super-resolution fluorescence imaging allows the visualization of the cellular morphology and interactions at the nanoscale. Statistical analysis methods such as super-resolution optical fluctuation imaging (SOFI) obtain an improved spatial resolution by analyzing fluorophore blinking but can be perturbed by the presence of non-stationary processes such as photodestruction or fluctuations in the illumination. In this work, we propose to use Whittaker smoothing to remove these smooth signal trends and retain only the information associated to independent blinking of the emitters, thus enhancing the SOFI signals. We find that our method works well to correct photodestruction, especially when it occurs quickly. The resulting images show a much higher contrast, strongly suppressed background and a more detailed visualization of cellular structures. Our method is parameter-free and computationally efficient, and can be readily applied on both two-dimensional and three-dimensional data.

Automated Analysis of Single-Molecule Photobleaching Data by Statistical Modeling of Spot Populations.

The number of fluorophores within a molecule complex can be revealed by single-molecule photobleaching imaging. A widely applied strategy to analyze intensity traces over time is the quantification of photobleaching step counts. However, several factors can limit and bias the detection of photobleaching steps, including noise, high numbers of fluorophores, and the possibility that several photobleaching events occur almost simultaneously. In this study, we propose a new approach, to our knowledge, to determine the fluorophore number that correlates the intensity decay of a population of molecule complexes with the decay of the number of visible complexes. We validated our approach using single and fourfold Atto-labeled DNA strands. As an example we estimated the subunit stoichiometry of soluble CD95L using GFP fusion proteins. To assess the precision of our method we performed in silico experiments showing that the estimates are not biased for experimentally observed intensity fluctuations and that the relative precision remains constant with increasing number of fluorophores. In case of fractional fluorescent labeling, our simulations predicted that the fluorophore number estimate corresponds to the product of the true fluorophore number with the labeling fraction. Our method, denoted by spot number and intensity correlation (SONIC), is fully automated, robust to noise, and does not require the counting of photobleaching events.

Differentiation between Shallow and Deep Charge Trap States on Single Poly(3-hexylthiophene) Chains through Fluorescence Photon Statistics.

Blinking of the photoluminescence (PL) emitted from individual conjugated polymer chains is one of the central observations made by single-molecule spectroscopy (SMS). Important information, for example regarding excitation energy transfer, can be extracted by evaluating dynamic quenching. However, the nature of trap states, which are responsible for PL quenching, often remains obscured. We present a detailed investigation of the photon statistics of single poly(3-hexylthiophene) (P3HT) chains obtained by SMS. The photon statistics provide a measure of the number and brightness of independently emitting areas on a single chain. These observables can be followed during blinking. A decrease in PL intensity is shown to be correlated with either 1) a decrease in the average brightness of the emitting sites; or 2) a decrease in the number of emitting regions. We attribute these phenomena to the formation of 1) shallow charge traps, which can weakly affect all emitting areas of a single chain at once; and 2) deep traps, which have a strong effect on small regions within the single chains.

Single-molecule studies on the label number distribution of fluorescent markers.

Over the past decade, a vast variety of different fluorescent labeling systems have emerged for use in fluorescence microscopy and fluorescence-based analytical techniques. A difficulty frequently arising when quantifying fluorescently labeled samples is that the number of labels per protein is neither well defined, for example, due to multiple functional groups that can undergo covalent coupling with activated dyes, nor well known, for example, due to limited methods mostly estimating ensemble averages. Herein, we use a recently established method that evaluates the statistics of multiple photon detection events to measure the label number distribution of different fluorescent marker molecules at the single-molecule level. We tested five different far-red dyes frequently used for fluorescence labeling and found all of them suitable for our counting method. We used two dyes, ATTO647N and Alexa647, to investigate the label number distribution of fluorescently labeled proteins. In the experiments, we found that the label number distribution of antibodies and streptavidin has a significant fraction of molecules labeled with two, three, or more fluorophores. In contrast, the distribution of label numbers for nanobodies resembles the one acquired for SNAP-tag, which can have a maximum of one label per protein. This is also reflected in the ensemble degree of labeling, which is in good agreement for the latter samples, whereas stronger deviations were observed for antibodies and streptavidin. Our single-molecule studies enable full characterization of the label number distribution for various fluorescent markers. This work puts quantitative studies on the stoichiometry of fluorescently tagged oligomers and protein aggregates into perspective.

Precise quantification of transcription factors in a surface-based single-molecule assay.

Biosensors have recognized a rapid development the last years in both industry and science. Recently, a single-molecule assay based on alternating laser excitation has been established for the quantitative detection of transcription factors. These proteins specifically recognize and bind DNA and play an important role in controlling gene expression. We implemented this assay format on a total internal reflection fluorescence microscope to detect transcription factors with immobilized single-molecule DNA biosensors. We quantify transcription factors via colocalization of the two halves of their binding site with immobilized single molecules of a two-color DNA biosensor. We could detect a model transcription factor, the bacterial lactose repressor, at different concentrations down to 150pM. We found that robust modeling of stoichiometry derived TIRF data is achieved with Student's t-distributions and nonlinear least-squares estimation with weights equal to the inverse of the expected number of bin entries. This significantly improved transcription factor concentration estimates with respect to distribution modeling with Gaussians without adding notable computational effort. The proposed model may enhance the precision of other single-molecule assays quantifying molecular distributions. Our measurements reliably confirm that the immobilized biosensor format is more sensitive than a previously published solution based approach.

Counting fluorescent dye molecules on DNA origami by means of photon statistics.

Obtaining quantitative information about molecular assemblies with high spatial and temporal resolution is a challenging task in fluorescence microscopy. Single-molecule techniques build on the ability to count molecules one by one. Here, a method is presented that extends recent approaches to analyze the statistics of coincidently emitted photons to enable reliable counting of molecules in the range of 1-20. This method does not require photochemistry such as blinking or bleaching. DNA origami structures are labeled with up to 36 dye molecules as a new evaluation tool to characterize this counting by a photon statistics approach. Labeled DNA origami has a well-defined labeling stoichiometry and ensures equal brightness for all dyes incorporated. Bias and precision of the estimating algorithm are determined, along with the minimal acquisition time required for robust estimation. Complexes containing up to 18 molecules can be investigated non-invasively within 150 ms. The method might become a quantifying add-on for confocal microscopes and could be especially powerful in combination with STED/RESOLFT-type microscopy.