Automated Stoichiometry Analysis of Single-Molecule Fluorescence Imaging Traces via Deep Learning

Dual-Label Single-Molecule Imaging Method for Quantifying Apparent Fluorescence Efficiency of Fluorescent Proteins

Fluorescent proteins (FPs) are indispensable tools for life science research, crucial for quantitative analyses, such as protein stoichiometry determination, signal transduction, and protein-protein interactions. In this work, we introduce a new method of dual-color single-molecule imaging to assess the apparent fluorescence efficiency of FPs (DC-FEFP). By integrating high signal-to-noise ratio (SNR) FPs or self-labeling tags, this approach enables us to precisely quantify the fluorescence efficiency of various FPs at the single-molecule level in both living and fixed cells. With DC-FEFP, we found that mNeonGreen has the highest fluorescence efficiency among three commonly used FPs in living cells, as well as the significant impact of fixation on the photophysical properties of FPs. DC-FEFP is a high-precision and versatile tool for quantifying the fluorescence efficiency of FPs. It is capable of providing accurate information to calibrate protein stoichiometry based on single-molecule imaging.

A Chemistry-Informed Generative Deep Learning Approach for Enhancing Voltammetric Neurochemical Sensing in Living Mouse Brain

Exploring the time-resolved dynamics of neurochemicals is essential for deciphering neuronal functions, intercellular communication, and neurophysiological or pathological mechanisms. However, the complex interplay among neurochemicals between neurocytes, coupled with extensive chemical signal crosstalk, puts simultaneous sensing of multiple neurochemicals into a longstanding challenge. Herein, we report a chemistry-informed generative neural network (CIGNN) model to separate the Faradaic and the non-Faradaic components from voltammetric currents, minimizing their mutual interference and enhancing quantitative accuracy. With the assistance of generative deep learning, we successfully establish a new platform for in vivo neurochemical sensing, which is validated by simultaneously monitoring the dynamics of dopamine (DA), ascorbic acid (AA), and ionic strength in a neuroinflammation mouse model. We observe that the stimulation with KCl solution triggers a significant enhancement of AA efflux on the model mice (300 ± 50 μM) compared with that from the control mice (170 ± 20 μM), as well as a significant decrease of ion influx (55 ± 7 mM) compared with that from the control mice (120 ± 16 mM), while not evoking a significant change in the DA release from the model mice (2.8 ± 0.3 μM) versus that from the control mice (3.0 ± 0.5 μM). This work provides a robust tool for studying multineurochemical signaling and elucidating the molecular mechanisms underlying various brain activities.

Pretrained Deep Neural Network Kin-SiM for Single-Molecule FRET Trace Idealization

Single-molecule fluorescence resonance energy transfer (smFRET) has emerged as a pivotal technique for probing biomolecular dynamics over time at nanometer scales. Quantitative analyses of smFRET time traces remain challenging due to confounding factors such as low signal-to-noise ratios, photophysical effects such as bleaching and blinking, and the complexity of modeling the underlying biomolecular states and kinetics. The dynamic distance information shaping the smFRET trace powerfully uncovers even transient conformational changes in single biomolecules both at or far from equilibrium, relying on trace idealization to identify specific interconverting states. Conventional trace idealization methods based on hidden Markov models (HMMs) require substantial a priori knowledge of the system under study, manual intervention, and assumptions about the number of states and transition probabilities. Here, we present a deep learning framework using long short-term memory (LSTM) to automate the trace idealization, termed Kin-SiM. Our approach employs neural networks pretrained on simulated data to learn high-order correlations in the multidimensional FRET trajectories. Without user input of Markovian assumptions, the trained LSTM networks directly idealize the FRET traces to extract the number of underlying biomolecular states, their interstate dynamics, and associated kinetic parameters. On benchmark smFRET data sets, Kin-SiM achieves a performance similar to conventional HMM-based methods but with less hands-on time and lower risk of bias. We further systematically evaluate the key training factors that affect network performance to define the correct hyperparameter tuning for applying deep neural networks to smFRET data analyses.

FluoroTensor: Identification and tracking of colocalised molecules and their stoichiometries in multi-colour single molecule imaging via deep learning

The identification of photobleaching steps in single molecule fluorescence imaging is a well-established procedure for analysing the stoichiometries of molecular complexes. Nonetheless, the method is challenging with protein fluorophores because of the high levels of noise, rapid bleaching and highly variable signal intensities, all of which complicate methods based on statistical analyses of intensities to identify bleaching steps. It has recently been shown that deep learning by convolutional neural networks can yield an accurate analysis with a relatively short computational time. We describe here an improved use of such an approach that detects bleaching events even in the first time point of observation, and we have included this within an integrated software package incorporating fluorescence spot detection, colocalisation, tracking, FRET and photobleaching step analyses of single molecules or complexes. This package, known as FluoroTensor, is written in Python with a self-explanatory user interface.

Probing and imaging phospholipid dynamics in live cells

Distinct phospholipid species display specific distribution patterns across cellular membranes, which are important for their structural and signaling roles and for preserving the integrity and functionality of the plasma membrane and organelles. Recent advancements in lipid biosensor technology and imaging modalities now allow for direct observation of phospholipid distribution, trafficking, and dynamics in living cells. These innovations have markedly advanced our understanding of phospholipid function and regulation at both cellular and subcellular levels. Herein, we summarize the latest developments in phospholipid biosensor design and application, emphasizing the contribution of cutting-edge imaging techniques to elucidating phospholipid dynamics and distribution with unparalleled spatiotemporal precision.

Single Effective Complex Loading into Zero-Mode Waveguides Optimized with Fluorescence Evaluation at Quenching and Accumulation Checkpoints

Single-molecule detection with high accuracy and specialty plays an important role in biomedical diagnosis and screening. Zero-mode waveguides (ZMWs) enable the possibility of single biological molecule detection in real time. Nevertheless, the absence of a reliable assessment for single effective complex loading has constrained further applications of ZMWs in complex interaction. Both the quantity and activity of the complex loaded into ZMWs have a critical effect on the efficiency of detection. Herein, a fluorescence evaluation at quenching and accumulation checkpoints was established to assess and optimize single effective complex loading into ZMWs. A primer-template-enzyme ternary complex was designed, and then an evaluation for quantity statistics at the quenching checkpoint and functional activity at the accumulation checkpoint was used to validate the effectiveness of complexes loaded into ZMWs. By optimizing the parameters such as loading time, procedures, and enzyme amount, the single-molecule effective occupancy was increased to 25.48%, achieving 68.86% of the theoretical maximum value (37%) according to Poisson statistics. It is of great significance to provide effective complex-loading validation for improving the sample-loading efficiency of single-molecule assays or sequencing in the future.

Deep Learning-Assisted Spectrum-Structure Correlation: State-of-the-Art and Perspectives

Spectrum-structure correlation is playing an increasingly crucial role in spectral analysis and has undergone significant development in recent decades. With the advancement of spectrometers, the high-throughput detection triggers the explosive growth of spectral data, and the research extension from small molecules to biomolecules accompanies massive chemical space. Facing the evolving landscape of spectrum-structure correlation, conventional chemometrics becomes ill-equipped, and deep learning assisted chemometrics rapidly emerges as a flourishing approach with superior ability of extracting latent features and making precise predictions. In this review, the molecular and spectral representations and fundamental knowledge of deep learning are first introduced. We then summarize the development of how deep learning assist to establish the correlation between spectrum and molecular structure in the recent 5 years, by empowering spectral prediction (i.e., forward structure-spectrum correlation) and further enabling library matching and de novo molecular generation (i.e., inverse spectrum-structure correlation). Finally, we highlight the most important open issues persisted with corresponding potential solutions. With the fast development of deep learning, it is expected to see ultimate solution of establishing spectrum-structure correlation soon, which would trigger substantial development of various disciplines.

Single-Molecule Stoichiometry of Supramolecular Complexes

The use of single-molecule microscopy is introduced as a method to quantify the photophysical properties of supramolecular complexes rapidly at ultra low concentrations (<1 nM), previously inaccessible. Using a model supramolecular system based on the host-guest complexation of cucurbit[n]uril (CB[n]) macrocycles together with a fluorescent guest (Ant910Me), we probe fluorescent CB[n] host-guest complexes in the single molecule regime. We show quantification and differentiation of host-guest photophysics and stoichiometries, both in aqueous media and noninvasively in hydrogel, by thresholding detected photons. This methodology has wide reaching implications in aiding the design of next-generation materials with programmed and controlled properties.

Construction of single-molecule counting-based biosensors for DNA-modifying enzymes: A review

DNA-modifying enzymes act as critical regulators in a wide range of genetic functions (e.g., DNA damage & repair, DNA replication), and their aberrant expression may interfere with regular genetic functions and induce various malignant diseases including cancers. DNA-modifying enzymes have emerged as the potential biomarkers in early diagnosis of diseases and new therapeutic targets in genomic research. Consequently, the development of highly specific and sensitive biosensors for the detection of DNA-modifying enzymes is of great importance for basic biomedical research, disease diagnosis, and drug discovery. Single-molecule fluorescence detection has been widely implemented in the field of molecular diagnosis due to its simplicity, high sensitivity, visualization capability, and low sample consumption. In this paper, we summarize the recent advances in single-molecule counting-based biosensors for DNA-modifying enzyme (i.e, alkaline phosphatase, DNA methyltransferase, DNA glycosylase, flap endonuclease 1, and telomerase) assays in the past four years (2019 - 2023). We highlight the principles and applications of these biosensors, and give new insight into the future challenges and perspectives in the development of single-molecule counting-based biosensors.

Deep Learning Promotes Profiling of Multiple miRNAs in Single Extracellular Vesicles for Cancer Diagnosis

Extracellular vesicle microRNAs (EV miRNAs) are critical noninvasive biomarkers for early cancer diagnosis. However, accurate cancer diagnosis based on bulk analysis is hindered by the heterogeneity among EVs. Herein, we report an approach for profiling single-EV multi-miRNA signatures by combining total internal reflection fluorescence (TIRF) imaging with a deep learning (DL) algorithm for the first time. This innovative technique allows for the precise characterization of EV miRNAs at the single-vesicle level, overcoming the challenges posed by EV heterogeneity. TIRF with high resolution and a signal-to-noise ratio can simultaneously detect multi-miRNAs in situ in individual EVs. DL algorithm avoids complicated and inaccurate artificial feature extraction, achieving automated high-resolution image analysis. Using this approach, we reveal that the main variation of EVs from 5 cancer cells and normal plasma is the triple-positive EV subpopulation, and the classification accuracy of single triple-positive EVs from 6 sources can reach above 95%. In the clinical cohort, 20 patients (5 lung cancer, 5 breast cancer, 5 cervical cancer, and 5 colon cancer) and 5 healthy controls are predicted with an overall accuracy of 100%. This single-EV strategy provides new opportunities for exploring more specific EV biomarkers to achieve cancer diagnosis and classification.

Deep Learning-Assisted Automated Multidimensional Single Particle Tracking in Living Cells

The translational and rotational dynamics of anisotropic optical nanoprobes revealed in single particle tracking (SPT) experiments offer molecular-level information about cellular activities. Here, we report an automated high-speed multidimensional SPT system integrated with a deep learning algorithm for tracking the 3D orientation of anisotropic gold nanoparticle probes in living cells with high localization precision (<10 nm) and temporal resolution (0.9 ms), overcoming the limitations of rotational tracking under low signal-to-noise ratio (S/N) conditions. This method can resolve the azimuth (0°-360°) and polar angles (0°-90°) with errors of less than 2° on the experimental and simulated data under S/N of ∼4. Even when the S/N approaches the limit of 1, this method still maintains better robustness and noise resistance than the conventional pattern matching methods. The usefulness of this multidimensional SPT system has been demonstrated with a study of the motions of cargos transported along the microtubules within living cells.

Holistic Prediction of AuNP Aggregation in Diverse Aqueous Suspensions Based on Machine Vision and Dark-Field Scattering Imaging

The localized surface-plasmon resonance of the AuNP in aqueous media is extremely sensitive to environmental changes. By measuring the signal of plasmon scattering light, the dark-field microscopic (DFM) imaging technique has been used to monitor the aggregation of AuNPs, which has attracted great attention because of its simplicity, low cost, high sensitivity, and universal applicability. However, it is still challenging to interpret DFM images of AuNP aggregation due to the heterogeneous characteristics of the isolated and discontinuous color distribution. Herein, we introduce machine vision algorithms for the training of DFM images of AuNPs in different saline aqueous media. A visual deep learning framework based on AlexNet is constructed for studying the aggregation patterns of AuNPs in aqueous suspensions, which allows for rapid and accurate identification of the aggregation extent of AuNPs, with a prediction accuracy higher than 0.96. With the aid of machine learning analysis, we further demonstrate the prediction ability of various aggregation phenomena induced by both cation species and the concentration of the external saline solution. Our results suggest the great potential of machine vision frameworks in the accurate recognition of subtle pattern changes in DFM images, which can help researchers build predictive analytics based on DFM imaging data.

Single-Molecule Imaging Reveals Differential AT1R Stoichiometry Change in Biased Signaling

G protein-coupled receptors (GPCRs) represent promising therapeutic targets due to their involvement in numerous physiological processes mediated by downstream G protein- and β-arrestin-mediated signal transduction cascades. Although the precise control of GPCR signaling pathways is therapeutically valuable, the molecular details for governing biased GPCR signaling remain elusive. The Angiotensin II type 1 receptor (AT1R), a prototypical class A GPCR with profound implications for cardiovascular functions, has become a focal point for biased ligand-based clinical interventions. Herein, we used single-molecule live-cell imaging techniques to evaluate the changes in stoichiometry and dynamics of AT1R with distinct biased ligand stimulations in real time. It was revealed that AT1R existed predominantly in monomers and dimers and underwent oligomerization upon ligand stimulation. Notably, β-arrestin-biased ligands induced the formation of higher-order aggregates, resulting in a slower diffusion profile for AT1R compared to G protein-biased ligands. Furthermore, we demonstrated that the augmented aggregation of AT1R, triggered by activation from each biased ligand, was completely abrogated in β-arrestin knockout cells. These findings furnish novel insights into the intricate relationship between GPCR aggregation states and biased signaling, underscoring the pivotal role of molecular behaviors in guiding the development of selective therapeutic agents.

Deep-LASI: deep-learning assisted, single-molecule imaging analysis of multi-color DNA origami structures

Single-molecule experiments have changed the way we explore the physical world, yet data analysis remains time-consuming and prone to human bias. Here, we introduce Deep-LASI (Deep-Learning Assisted Single-molecule Imaging analysis), a software suite powered by deep neural networks to rapidly analyze single-, two- and three-color single-molecule data, especially from single-molecule Förster Resonance Energy Transfer (smFRET) experiments. Deep-LASI automatically sorts recorded traces, determines FRET correction factors and classifies the state transitions of dynamic traces all in ~20-100 ms per trajectory. We benchmarked Deep-LASI using ground truth simulations as well as experimental data analyzed manually by an expert user and compared the results with a conventional Hidden Markov Model analysis. We illustrate the capabilities of the technique using a highly tunable L-shaped DNA origami structure and use Deep-LASI to perform titrations, analyze protein conformational dynamics and demonstrate its versatility for analyzing both total internal reflection fluorescence microscopy and confocal smFRET data.

Single-molecule tracking of Nanog and Oct4 in living mouse embryonic stem cells uncovers a feedback mechanism of pluripotency maintenance

Nanog and Oct4 are core transcription factors that form part of a gene regulatory network to regulate hundreds of target genes for pluripotency maintenance in mouse embryonic stem cells (ESCs). To understand their function in the pluripotency maintenance, we visualised and quantified the dynamics of single molecules of Nanog and Oct4 in a mouse ESCs during pluripotency loss. Interestingly, Nanog interacted longer with its target loci upon reduced expression or at the onset of differentiation, suggesting a feedback mechanism to maintain the pluripotent state. The expression level and interaction time of Nanog and Oct4 correlate with their fluctuation and interaction frequency, respectively, which in turn depend on the ESC differentiation status. The DNA viscoelasticity near the Oct4 target locus remained flexible during differentiation, supporting its role either in chromatin opening or a preferred binding to uncondensed chromatin regions. Based on these results, we propose a new negative feedback mechanism for pluripotency maintenance via the DNA condensation state-dependent interplay of Nanog and Oct4.

Discovering the digital biomarker of hepatocellular carcinoma in serum with SERS-based biosensors and intelligence vision

By its many virtues, non-biomarker-reliant molecular detection has recently shown bright prospects for cancer screening but its clinical application is hindered by the shortage of measurable criteria that are analogous to biomarkers. Here, we report a digital biomarker, as a new-concept serum biomarker, of hepatocellular carcinoma (HCC) found with SERS-based biosensors and a deep neural network "digital retina" for visualizing and explicitly defining spectral fingerprints. We validate the discovered digital biomarker (a collection of 10 characteristic peaks in the serum SERS spectra) with unsupervised clustering of spectra from an independent sample batch comprised normal individuals and HCC cases; the validation results show clustering accuracies of 95.71% and 100.00%, respectively. Furthermore, we find that the digital biomarker of HCC shares a few common peaks with three clinically applied serum biomarkers, which means it could convey essential biomolecular information similar to these biomarkers. Accordingly, we present an intelligent method for early HCC detection that leverages the digital biomarker with similar traits as biomarkers. Employing the digital biomarker, we could accurately stratify HCC, hepatitis B, and normal populations with linear classifiers, exhibiting accuracies over 92% and area under the receiver operating curve values above 0.93. It is anticipated that this non-biomarker-reliant molecular detection method will facilitate mass cancer screening.

Deep Learning Enhanced Electrochemiluminescence Microscopy

Limited by the efficiency of electrochemiluminescence, tens of seconds of exposure time are typically required to get a high-quality image. Image enhancement of short exposure time images to obtain a well-defined electrochemiluminescence image can meet the needs of high-throughput or dynamic imaging. Here, we propose deep enhanced ECL microscopy (DEECL), a general strategy that utilizes artificial neural networks to reconstruct electrochemiluminescence images with millisecond exposure times to have similar quality as high-quality electrochemiluminescence images with second-long exposure time. Electrochemiluminescence imaging of fixed cells demonstrates that DEECL allows improvement of the imaging efficiency by 1 to 2 orders than usual. This approach is further used for a data-intensive analysis application, cell classification, achieving an accuracy of 85% with ECL data at an exposure time of 50 ms. We anticipate that the computationally enhanced electrochemiluminescence microscopy will enable fast and information-rich imaging and prove useful for understanding dynamic chemical and biological processes.

A near infrared two-channel fluorescent probe for the detection of hydrogen sulfide and viscosity with a negligible crosstalk influence

Both imbalance of H₂S production and the change of viscosity in cells are associated with many diseases such as inflammation, Alzheimer's disease, and Parkinson's disease. Thus, the development of two-channel fluorescent probes for the detection of H₂S and viscosity is of great significance for the study of pathogenic mechanisms. Herein, we design a two-channel NIR fluorescent probe RHO-DCO-DNP, which was able to selectively respond to H₂S in one channel (λex = 580 nm, λem = 760 nm) and to viscosity in another channel (λex = 400 nm, λem = 585 nm). It should be emphasized that there is a negligible impact from the crosstalk between the two optical channels and the two targets. In addition, with the low cytotoxicity and unique dual lysosome/mitochondria targeting capability, the probe was successfully applied to the sensing of H₂S and viscosity in normal cells and inflammation cells through fluorescent imaging. The probe could be a promising molecular tool for exploring the pathological role of H₂S, viscosity, and both of them.

Pore-Forming Proteins: From Pore Assembly to Structure by Quantitative Single-Molecule Imaging

Pore-forming proteins (PFPs) play a central role in many biological processes related to infection, immunity, cancer, and neurodegeneration. A common feature of PFPs is their ability to form pores that disrupt the membrane permeability barrier and ion homeostasis and generally induce cell death. Some PFPs are part of the genetically encoded machinery of eukaryotic cells that are activated against infection by pathogens or in physiological programs to carry out regulated cell death. PFPs organize into supramolecular transmembrane complexes that perforate membranes through a multistep process involving membrane insertion, protein oligomerization, and finally pore formation. However, the exact mechanism of pore formation varies from PFP to PFP, resulting in different pore structures with different functionalities. Here, we review recent insights into the molecular mechanisms by which PFPs permeabilize membranes and recent methodological advances in their characterization in artificial and cellular membranes. In particular, we focus on single-molecule imaging techniques as powerful tools to unravel the molecular mechanistic details of pore assembly that are often obscured by ensemble measurements, and to determine pore structure and functionality. Uncovering the mechanistic elements of pore formation is critical for understanding the physiological role of PFPs and developing therapeutic approaches.

DNA-Programed Plasmon Rulers Decrypt Single-Receptor Dimerization on Cell Membrane

Decrypting the dynamics of receptor dimerization on cell membranes bears great importance in identifying the mechanisms regulating diverse cellular activities. In this regard, long-term monitoring of single-molecule behavior during receptor dimerization allows deepening insight into the dimerization process and tracking of the behavior of individual receptors, yet this remains to be realized. Herein, real-time observation of the receptor tyrosine kinases family (RTKs) at single-molecule level based on plasmon rulers was achieved for the first time, which enabled precise regulation and dynamic monitoring of the dimerization process by DNA programming with excellent photostability. Additionally, those nanoprobes demonstrated substantial application in the regulation of RTKs protein dimerization/phosphorylation and activation of downstream signaling pathways. The proposed nanoprobes hold considerable potential for elucidating the molecular mechanisms of single-receptor dimerization as well as the conformational transitions upon dimerization, providing a new paradigm for the precise manipulation and monitoring of specific single-receptor crosslink events in biological systems.

Feature Selection Assists BLSTM for the Ultrasensitive Detection of Bioflavonoids in Different Biological Matrices Based on the 3D Fluorescence Spectra of Gold Nanoclusters

Rapid and on-site qualitative and quantitative analysis of small molecules (including bioflavonoids) in biofluids are of great importance in biomedical applications. Herein, we have developed two deep learning models based on the 3D fluorescence spectra of gold nanoclusters as a single probe for rapid qualitative and quantitative analysis of eight bioflavonoids in serum. The results proved the efficiency and stability of the random forest-bidirectional long short-term memory (RF-BLSTM) model, which was used only with the most important features after deleting the unimportant features that might hinder the performance of the model in identifying the selected bioflavonoids in serum at very low concentrations. The optimized model achieves excellent overall accuracy (98-100%) in the qualitative analysis of the selected bioflavonoids. Next, the optimized model was transferred to quantify the selected bioflavonoids in serum at nanoscale concentrations. The transferred model achieved excellent accuracy, and the overall determination coefficient (R²) value range was 99-100%. Furthermore, the optimized model achieved excellent accuracies in other applications, including multiplex detection in serum and model applicability in urine. Also, LOD in serum at nanoscale concentration was considered. Therefore, this approach opens the window for qualitative and quantitative analysis of small molecules in biofluids at nanoscale concentrations, which may help in the rapid inclusion of sensor arrays in biomedical and other applications.

Nanomolar LL-37 induces permeability of a biomimetic mitochondrial membrane

LL-37, the only human host cathelicidin peptide, is proposed to be able to induce host cell apoptosis through mitochondrial membrane permeabilization (MMP). Detailed pathways of the LL-37-triggered MMP are however still disputed. It is generally believed that cationic peptides permeate a membrane mostly in conditions of micromolar peptide concentrations and negatively charged membranes, which are not usually satisfied in the mitochondrial circumstance. Herein, using a variety of single-molecule techniques, we show that nanomolar LL-37 specifically induces permeability of a phosphoethanolamine (PE)-rich biomimetic mitochondrial membrane in a protein-independent manner. The insertion dynamics of single LL-37 molecules exhibit different metastable states in bilayers composed of different lipids. Moreover, the PE lipids significantly facilitate adsorption and accumulation of LL-37 on the PE-rich bilayer, and produce deeper insertion of peptide oligomers, especially tetramers, into the bilayer. This work offers an alternative pathway of the LL-37-triggered MMP and apoptosis.

Deep learning in single-molecule imaging and analysis: recent advances and prospects

Single-molecule microscopy is advantageous in characterizing heterogeneous dynamics at the molecular level. However, there are several challenges that currently hinder the wide application of single molecule imaging in bio-chemical studies, including how to perform single-molecule measurements efficiently with minimal run-to-run variations, how to analyze weak single-molecule signals efficiently and accurately without the influence of human bias, and how to extract complete information about dynamics of interest from single-molecule data. As a new class of computer algorithms that simulate the human brain to extract data features, deep learning networks excel in task parallelism and model generalization, and are well-suited for handling nonlinear functions and extracting weak features, which provide a promising approach for single-molecule experiment automation and data processing. In this perspective, we will highlight recent advances in the application of deep learning to single-molecule studies, discuss how deep learning has been used to address the challenges in the field as well as the pitfalls of existing applications, and outline the directions for future development.

Toward high-throughput oligomer detection and classification for early-stage aggregation of amyloidogenic protein

Aggregation kinetics of proteins and peptides have been studied extensively due to their significance in many human diseases, including neurodegenerative disorders, and the roles they play in some key physiological processes. However, most of these studies have been performed as bulk measurements using Thioflavin T or other fluorescence turn-on reagents as indicators of fibrillization. Such techniques are highly successful in making inferences about the nucleation and growth mechanism of fibrils, yet cannot directly measure assembly reactions at low protein concentrations which is the case for amyloid-β (Aβ) peptide under physiological conditions. In particular, the evolution from monomer to low-order oligomer in early stages of aggregation cannot be detected. Single-molecule methods allow direct access to such fundamental information. We developed a high-throughput protocol for single-molecule photobleaching experiments using an automated fluorescence microscope. Stepwise photobleaching analysis of the time profiles of individual foci allowed us to determine stoichiometry of protein oligomers and probe protein aggregation kinetics. Furthermore, we investigated the potential application of supervised machine learning with support vector machines (SVMs) as well as multilayer perceptron (MLP) artificial neural networks to classify bleaching traces into stoichiometric categories based on an ensemble of measurable quantities derivable from individual traces. Both SVM and MLP models achieved a comparable accuracy of more than 80% against simulated traces up to 19-mer, although MLP offered considerable speed advantages, thus making it suitable for application to high-throughput experimental data. We used our high-throughput method to study the aggregation of Aβ40 in the presence of metal ions and the aggregation of α-synuclein in the presence of gold nanoparticles.

Deep Learning-Based Sensor Array: 3D Fluorescence Spectra of Gold Nanoclusters for Qualitative and Quantitative Analysis of Vitamin B6 Derivatives

Vitamin B₆ derivatives (VB6Ds) are of great importance for all living organisms to complete their physiological processes. However, their excess in the body can cause serious problems. What is more, the qualitative and quantitative analysis of different VB6Ds may present significant challenges due to the high similarity of their chemical structures. Also, the transfer of deep learning model from one task to a similar task needs to be present more in the fluorescence-based biosensor. Therefore, to address these problems, two deep learning models based on the intrinsic fingerprint of 3D fluorescence spectra have been developed to identify five VB6Ds. The accuracy ranges of a deep neural network (DNN) and a convolutional neural network (CNN) were 94.44-97.77% and 97.77-100%, respectively. After that, the developed models were transferred for quantitative analysis of the selected VB6Ds at a broad concentration range (1-100 μM). The determination coefficient (R²) values of the test set for DNN and CNN were 93.28 and 97.01%, respectively, which also represents the outperformance of CNN over DNN. Therefore, our approach opens new avenues for qualitative and quantitative sensing of small molecules, which will enrich fields related to deep learning, analytical chemistry, and especially sensor array chemistry.

Carbon Dot Blinking Fingerprint Uncovers Native Membrane Receptor Organizations via Deep Learning

Oligomeric organization of G protein-coupled receptors is proposed to regulate receptor signaling and function, yet rapid and precise identification of the oligomeric status especially for native receptors on a cell membrane remains an outstanding challenge. By using blinking carbon dots (CDs), we now develop a deep learning (DL)-based blinking fingerprint recognition method, named deep-blinking fingerprint recognition (BFR), which allows automatic classification of CD-labeled receptor organizations on a cell membrane. This DL model integrates convolutional layers, long-short-term memory, and fully connected layers to extract time-dependent blinking features of CDs and is trained to a high accuracy (∼95%) for identifying receptor organizations. Using deep blinking fingerprint recognition, we found that CXCR4 mainly exists as 87.3% monomers, 12.4% dimers, and <1% higher-order oligomers on a HeLa cell membrane. We further demonstrate that the heterogeneous organizations can be regulated by various stimuli at different degrees. The receptor-binding ligands, agonist SDF-1α and antagonist AMD3100, can induce the dimerization of CXCR4 to 33.1 and 20.3%, respectively. In addition, cytochalasin D, which inhibits actin polymerization, similarly prompts significant dimerization of CXCR4 to 30.9%. The multi-pathway organization regulation will provide an insight for understanding the oligomerization mechanism of CXCR4 as well as for elucidating their physiological functions.

Diffraction-Limited Molecular Cluster Quantification with Bayesian Nonparametrics

Life's fundamental processes involve multiple molecules operating in close proximity within cells. To probe the composition and kinetics of molecular clusters confined within small (diffraction-limited) regions, experiments often report on the total fluorescence intensity simultaneously emitted from labeled molecules confined to such regions. Methods exist to enumerate total fluorophore numbers (e.g., step counting by photobleaching). However, methods aimed at step counting by photobleaching cannot treat photophysical dynamics in counting nor learn their associated kinetic rates. Here we propose a method to simultaneously enumerate fluorophores and determine their individual photophysical state trajectories. As the number of active (fluorescent) molecules at any given time is unknown, we rely on Bayesian nonparametrics and use specialized Monte Carlo algorithms to derive our estimates. Our formulation is benchmarked on synthetic and real data sets. While our focus here is on photophysical dynamics (in which labels transition between active and inactive states), such dynamics can also serve as a proxy for other types of dynamics such as assembly and disassembly kinetics of clusters. Similarly, while we focus on the case where all labels are initially fluorescent, other regimes, more appropriate to photoactivated localization microscopy, where fluorophores are instantiated in a non-fluorescent state, fall within the scope of the framework. As such, we provide a complete and versatile framework for the interpretation of complex time traces arising from the simultaneous activity of up to 100 fluorophores.

Systematic Assessment of the Accuracy of Subunit Counting in Biomolecular Complexes Using Automated Single-Molecule Brightness Analysis

Analysis of single-molecule brightness allows subunit counting of high-order oligomeric biomolecular complexes. Although the theory behind the method has been extensively assessed, systematic analysis of the experimental conditions required to accurately quantify the stoichiometry of biological complexes remains challenging. In this work, we develop a high-throughput, automated computational pipeline for single-molecule brightness analysis that requires minimal human input. We use this strategy to systematically quantify the accuracy of counting under a wide range of experimental conditions in simulated ground-truth data and then validate its use on experimentally obtained data. Our approach defines a set of conditions under which subunit counting by brightness analysis is designed to work optimally and helps in establishing the experimental limits in quantifying the number of subunits in a complex of interest. Finally, we combine these features into a powerful, yet simple, software that can be easily used for the analysis of the stoichiometry of such complexes.

Next generation single-molecule techniques: Imaging, labeling, and manipulation in vitro and in cellulo

Owing to their unique abilities to manipulate, label, and image individual molecules in vitro and in cellulo, single-molecule techniques provide previously unattainable access to elementary biological processes. In imaging, single-molecule fluorescence resonance energy transfer (smFRET) and protein-induced fluorescence enhancement in vitro can report on conformational changes and molecular interactions, single-molecule pull-down (SiMPull) can capture and analyze the composition and function of native protein complexes, and single-molecule tracking (SMT) in live cells reveals cellular structures and dynamics. In labeling, the abilities to specifically label genomic loci, mRNA, and nascent polypeptides in cells have uncovered chromosome organization and dynamics, transcription and translation dynamics, and gene expression regulation. In manipulation, optical tweezers, integration of single-molecule fluorescence with force measurements, and single-molecule force probes in live cells have transformed our mechanistic understanding of diverse biological processes, ranging from protein folding, nucleic acids-protein interactions to cell surface receptor function.

Single-Molecule Fluorescence Imaging Reveals GABAB Receptor Aggregation State Changes

The GABAB receptor is a typical G protein–coupled receptor, and its functional impairment is related to a variety of diseases. While the premise of GABAB receptor activation is the formation of heterodimers, the receptor also forms a tetramer on the cell membrane. Thus, it is important to study the effect of the GABAB receptor aggregation state on its activation and signaling. In this study, we have applied single-molecule photobleaching step counting and single-molecule tracking methods to investigate the formation and change of GABAB dimers and tetramers. A single-molecule stoichiometry assay of the wild-type and mutant receptors revealed the key sites on the interface of ligand-binding domains of the receptor for its dimerization. Moreover, we found that the receptor showed different aggregation behaviors at different conditions. Our results offered new evidence for a better understanding of the molecular basis for GABAB receptor aggregation and activation.

Photobleaching step analysis for robust determination of protein complex stoichiometries

The counting of discrete photobleaching steps in fluorescence microscopy is ideally suited to study protein complex stoichiometry in situ. The counting range of photobleaching step analysis has been significantly improved with more-sophisticated algorithms for step detection, albeit at an increasing computational cost and with the necessity for high-quality data. Here, we address concerns regarding robustness, automation, and experimental validation, optimizing both data acquisition and analysis. To make full use of the potential of photobleaching step analysis, we evaluate various labeling strategies with respect to their molecular brightness, photostability, and photoblinking. The developed analysis algorithm focuses on automation and computational efficiency. Moreover, we validate the developed methods with experimental data acquired on DNA origami labeled with defined fluorophore numbers, demonstrating counting of up to 35 fluorophores. Finally, we show the power of the combination of optimized trace acquisition and automated data analysis by counting labeled nucleoporin 107 in nuclear pore complexes of intact U2OS cells. The successful in situ application promotes this framework as a new resource enabling cell biologists to robustly determine the stoichiometries of molecular assemblies at the single-molecule level in an automated manner.

Deep Learning for Voltammetric Sensing in a Living Animal Brain

Numerous neurochemicals have been implicated in the modulation of brain function, making them appealing analytes for sensors and diagnostics. However, it is a grand challenge to selectively measure multiple neurochemicals simultaneously in vivo because of their great variations in concentrations, dynamic nature, and composition. Herein, we present a deep learning-based voltammetric sensing platform for the highly selective and simultaneous analysis of three neurochemicals in a living animal brain. The system features a carbon fiber electrode capable of capturing the mixed dynamics of a neurotransmitter, neuromodulator, and ions. Then a powerful deep neural network is employed to resolve individual chemical and spatial-temporal information. With this, a single electrochemical measurement reveals an interplaying concentration changes of dopamine, ascorbate, and ions in living rat brain, which is unobtainable with existing analytical methodologies. Our strategy provides a powerful means to expedite research in neuroscience and empower sensing-aided diagnostic applications.

Unsupervised selection of optimal single-molecule time series idealization criterion

Single-molecule (SM) approaches have provided valuable mechanistic information on many biophysical systems. As technological advances lead to ever-larger data sets, tools for rapid analysis and identification of molecules exhibiting the behavior of interest are increasingly important. In many cases the underlying mechanism is unknown, making unsupervised techniques desirable. The divisive segmentation and clustering (DISC) algorithm is one such unsupervised method that idealizes noisy SM time series much faster than computationally intensive approaches without sacrificing accuracy. However, DISC relies on a user-selected objective criterion (OC) to guide its estimation of the ideal time series. Here, we explore how different OCs affect DISC’s performance for data typical of SM fluorescence imaging experiments. We find that OCs differing in their penalty for model complexity each optimize DISC’s performance for time series with different properties such as signal/noise and number of sample points. Using a machine learning approach, we generate a decision boundary that allows unsupervised selection of OCs based on the input time series to maximize performance for different types of data. This is particularly relevant for SM fluorescence data sets, which often have signal/noise near the derived decision boundary and include time series of nonuniform length because of stochastic bleaching. Our approach, AutoDISC, allows unsupervised per-molecule optimization of DISC, which will substantially assist in the rapid analysis of high-throughput SM data sets with noisy samples and nonuniform time windows.

Deep-Learning-Assisted Single-Molecule Tracking on a Live Cell Membrane

Single-molecule fluorescence imaging is a powerful tool to study protein function by tracking molecular position and distribution, but the precise and rapid identification of dynamic molecules remains challenging due to the heterogeneous distribution and interaction of proteins on the live cell membrane. We now develop a deep-learning (DL)-assisted single-molecule imaging method that can precisely distinguish the monomer and complex for rapid and real-time tracking of protein interaction. This DL-based model, which comprises convolutional layers, max pooling layers, and fully connected layers, is trained to reach an accuracy of >98% for identifying monomer and complex. We use this method to investigate the dynamic process of chemokine receptor CXCR4 on the live cell membrane during the early signaling stage. The results show that, upon ligand activation, the CXCR4 undergoes a dynamic process of forming a receptor complex. We further demonstrate that the CXCR4 complex tends to be internalized at 2.5-fold higher rate into the cell interior than the monomer via the clathrin-dependent pathway. This study is the first example to scrutinize the early signaling process of CXCR4 at the single-molecule level on the live cell membrane. We envision that this DL-assisted imaging method would be a broadly useful technique to study more protein families for elucidating their physiological and pathological functions.

AutoSmarTrace: Automated chain tracing and flexibility analysis of biological filaments

Single-molecule imaging is widely used to determine statistical distributions of molecular properties. One such characteristic is the bending flexibility of biological filaments, which can be parameterized via the persistence length. Quantitative extraction of persistence length from images of individual filaments requires both the ability to trace the backbone of the chains in the images and sufficient chain statistics to accurately assess the persistence length. Chain tracing can be a tedious task, performed manually or using algorithms that require user input and/or supervision. Such interventions have the potential to introduce user-dependent bias into the chain selection and tracing. Here, we introduce a fully automated algorithm for chain tracing and determination of persistence lengths. Dubbed "AutoSmarTrace," the algorithm is built off a neural network, trained via machine learning to identify filaments within images recorded using atomic force microscopy. We validate the performance of AutoSmarTrace on simulated images with widely varying levels of noise, demonstrating its ability to return persistence lengths in agreement with input simulation parameters. Persistence lengths returned from analysis of experimental images of collagen and DNA agree with previous values obtained from these images with different chain-tracing approaches. Although trained on atomic-force-microscopy-like images, the algorithm also shows promise to identify chains in other single-molecule imaging approaches, such as rotary-shadowing electron microscopy and fluorescence imaging.

DeepSinse: deep learning based detection of single molecules

MOTIVATION

Imaging single molecules has emerged as a powerful characterization tool in the biological sciences. The detection of these under various noise conditions requires the use of algorithms that are dependent on the end-user inputting several parameters, the choice of which can be challenging and subjective.

RESULTS

In this work, we propose DeepSinse, an easily-trainable and useable deep neural network that can detect single molecules with little human input and across a wide range of signal-to-noise ratios. We validate the neural network on the detection of single bursts in simulated and experimental data and compare its performance with the best-in-class, domain-specific algorithms.

AVAILABILITY

Ground truth ROI simulating code, neural network training, validation code, classification code, ROI picker, GUI for simulating, training and validating DeepSinse as well as pre-trained networks are all released under the MIT License on www.github.com/jdanial/DeepSinse.The dSTORM dataset processing code is released under the MIT License on www.github.com/jdanial/StormProcessor.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Determining stoichiometry of ion channel complexes using single subunit counting

Most membrane proteins, and ion channels in particular, assemble to multimeric biological complexes. This starts with the quarternary structure and continues with the recruitment of auxiliary subunits and oligomerization or clustering of the complexes. While the quarternary structure is best determined by atomic-scale structures, stoichiometry of heteromers and dynamic changes in the assembly cannot necessarily be investigated with structural methods. Here, single subunit counting has proven a powerful method to study the composition of these complexes. Single subunit counting uses the irreversible photodestruction of fluorescent tags as means to directly count a labeled subunit and thereby derive the composition of the assemblies. In this chapter, we discuss single subunit counting and its limitations. We present alternative methods and provide a detailed protocol for recording and analysis of single subunit counting data.

An update on molecular counting in fluorescence microscopy

Quantitative assessment of protein complexes, such as receptor clusters in the context of cellular signalling, has become a pressing objective in cell biology. The advancements in the field of single molecule fluorescence microscopy have led to different approaches for counting protein copy numbers in various cellular structures. This has resulted in an increasing interest in robust calibration protocols addressing photophysical properties of fluorescent labels and the effect of labelling efficiencies. Here, we want to give an update on recent methods for protein counting with a focus on novel calibration protocols. In this context, we discuss different types of calibration samples and identify some of the challenges arising in molecular counting experiments. Some recently published applications offer potential approaches to tackle these challenges.

Deep Learning for Biospectroscopy and Biospectral Imaging: State-of-the-Art and Perspectives

With the advances in instrumentation and sampling techniques, there is an explosive growth of data from molecular and cellular samples. The call to extract more information from the large data sets has greatly challenged the conventional chemometrics method. Deep learning, which utilizes very large data sets for finding hidden features therein and for making accurate predictions for a wide range of applications, has been applied in an unbelievable pace in biospectroscopy and biospectral imaging in the recent 3 years. In this Feature, we first introduce the background and basic knowledge of deep learning. We then focus on the emerging applications of deep learning in the data preprocessing, feature detection, and modeling of the biological samples for spectral analysis and spectroscopic imaging. Finally, we highlight the challenges and limitations in deep learning and the outlook for future directions.

Automated Plasmonic Resonance Scattering Imaging Analysis via Deep Learning

Plasmonic nanoparticles, which have excellent local surface plasmon resonance (LSPR) optical and chemical properties, have been widely used in biology, chemistry, and photonics. The single-particle light scattering dark-field microscopy (DFM) imaging technique based on a color-coded analytical method is a promising approach for high-throughput plasmonic nanoparticle scatterometry. Due to the interference of high noise levels, accurately extracting real scattering light of plasmonic nanoparticles in living cells is still a challenging task, which hinders its application for intracellular analysis. Herein, we propose an automatic and high-throughput LSPR scatterometry technique using a U-Net convolutional deep learning neural network. We use the deep neural networks to recognize the scattering light of nanoparticles from background interference signals in living cells, which have a dynamic and complicated environment, and construct a DFM image semantic analytical model based on the U-Net convolutional neural network. Compared with traditional methods, this method can achieve higher accuracy, stronger generalization ability, and robustness. As a proof of concept, the change of intracellular cytochrome c in MCF-7 cells under UV light-induced apoptosis was monitored through the fast and high-throughput analysis of the plasmonic nanoparticle scattering light, providing a new strategy for scatterometry study and imaging analysis in chemistry.

An antibody-free liver cancer screening approach based on nanoplasmonics biosensing chips via spectrum-based deep learning

The clinical needs of rapidly screening liver cancer in large populations have asked for a facile and low-cost point-of-care testing (POCT) method. We present a nanoplasmonics biosensing chip (NBC) that would empower antibody-free detection with simplified analysis procedures for POCT. The cheaply fabricable NBC consists of multiple silver nanoparticle-decorated ZnO nanorods on cellulose filter paper and would enable one-drop blood tests through surface-enhanced Raman spectroscopy (SERS) detection. In this work, utilizing such an NBC and deep neural network (DNN) modeling, a direct serological detection platform was constructed for automatically identifying liver cancer within minutes. This chip could enhance Raman signals enough to be applied to POCT. A classification DNN model was established by spectrum-based deep learning with 1140 serum SERS spectra in equal proportions from hepatocellular carcinoma (HCC) patients and healthy individuals, achieving an identification accuracy of 91% on an external validation set of 100 spectra (50 HCC versus 50 healthy). The intelligent platform, based on the biosensing chip and DNN, has the potential for clinical applications and generalizable use in quickly screening or detecting other types of cancer.

The method for integrating dual-color fluorescence colocalization and single molecule photobleaching technology on the theophylline sensing platform

• Smart usage of single molecule photobleaching technology and dual-color fluorescence colocalization is of critical importance for exploiting the sensing platform. Here, we provide the detailed protocols related to the article “A split aptamer sensing platform for highly sensitive detection of theophylline based on dual-color fluorescence colocalization and single molecule photobleaching” (published online by Biosensors and Bioelectronics) (Liu et al., 2020). The protocols contain: (1) how to clean the slides; (2) how to prepare the probe and detection sample; (3) Single molecule imaging; 4) Data processing by using the Image J. Finally, we used a simple model to confirm the feasibility of the method for integrating dual-color fluorescence colocalization and single molecule photobleaching technology on the theophylline sensing platform.

• A simple, ultrasensitive method for the detection of theophylline.

• The method is easily comprehensible.

• Both strategy formulation and data processing are simple, learnability, and highly reproducible.

Analyzing protein dynamics from fluorescence intensity traces using unsupervised deep learning network

We propose an unsupervised deep learning network to analyze the dynamics of membrane proteins from the fluorescence intensity traces. This system was trained in an unsupervised manner with the raw experimental time traces and synthesized ones, so neither predefined state number nor pre-labelling were required. With the bidirectional Long Short-Term Memory (biLSTM) networks as the hidden layers, both the past and future context can be used fully to improve the prediction results and can even extract information from the noise distribution. The method was validated with the synthetic dataset and the experimental dataset of monomeric fluorophore Cy5, and then applied to extract the membrane protein interaction dynamics from experimental data successfully.

Flapping Peryleneimide as a Fluorogenic Dye with High Photostability and Strong Visible-Light Absorption

Flapping fluorophores (FLAP) with a flexible 8π ring are rapidly gaining attention as a versatile photofunctional system. Here we report a highly photostable "flapping peryleneimide" with an unprecedented fluorogenic mechanism based on a bent-to-planar conformational change in the S₁ excited state. The S₁ planarization induces an electronic configurational switch, almost quenching the inherent fluorescence (FL) of the peryleneimide moieties. However, the FL quantum yield is remarkably improved with a prolonged lifetime upon a slight environmental change. This fluorogenic function is realized by sensitive π-conjugation design, as a more π-expanded analogue does not show the planarization dynamics. With strong visible-light absorption, the FL lifetime response synchronized with the flexible flapping motion is useful for the latest optical techniques such as FL lifetime imaging microscopy (FLIM).

A split aptamer sensing platform for highly sensitive detection of theophylline based on dual-color fluorescence colocalization and single molecule photobleaching

A new split aptamer sensing platform is developed for highly sensitive and selective detection of theophylline based on single molecule photobleaching (SMPB) technique. The sensing system contains two probes. One is formed by one streptavidin and four biotinylated RNA fragments labelled with fluorescein isothiocyanate (FITC). Each biotinylated RNA fragment contains two repeating aptamer fragments. The other probe is the complementary aptamer fragment labelled with Cy5 dye. The existence of theophylline can trigger the first probe to bind as many as eight Cy5-labelled probes. The average combined number depends on the theophylline concentration and can be measured by SMPB technique. In the sensing system, the dual-color fluorescence colocalization is performed by the red fluorophore (Cy5) and green fluorophore (FITC), in which the red fluorophore is utilized for quantitative counting of photobleaching steps, while the green fluorophore serves as a counting reference to increase detection efficiency. On basis of the principle, an ultra-sensitive sensing platform of theophylline is created with a low limit of detection (LOD) of 0.092 nM. This work provides not only a highly sensitive method for theophylline detection but also a novel perspective for the applications of SMPB technology to construct biosensors.