"I firmly believe that markless will be accepted by the mainstream, and even surpass traditional testing methods in many ways. ”——Professor Chu Kaiqin of the University of Science and Technology of China said confidently when talking about the future direction of biochemical particle detection technology.
So what exactly is "unmarked"?
"Label-free" refers to the monitoring, parameter characterization, and related analysis of the physical and chemical properties of a chemical sample without fluorescent staining. It has the advantages of no light drift or phototoxicity, low requirements for sample preparation, and panoramic observation in time.
Compared with "immunoprecipitation" and "fluorescence staining", which are severely destructive or phototoxic to cell samples, "no labeling" can maximize the observation results for a long time without destroying the original state of the sample, and the results obtained in this way are closest to the performance of the sample in its own environment. Whether it is the observation of the activity of biological samples or the study of the mechanism of action of drugs, "no labeling" may be the most scientific and valuable way to refer to it in the future.
However, without labeling the sample, those small and faint particles in the sample are difficult to observe, especially in the state of rapid activity or agglomeration. These issues can hinder the application of "label-free" biochemical detection technology.
Professor Chu Kaiqin's team has been committed to the research of Raman signal analysis technology, image algorithm and other fields for many years, and has successfully broken this technical barrier in the industry, allowing "markerless" to serve the field of biochemical detection, and has achieved a series of excellent research results.
Here we must be clear about the concept of "Raman".
The Raman signal is a very weak inelastic scattering signal (only about 1 10 9 of the scattered light), compared with Rayleigh scattering (the scattered light has the same wavelength as the incident light), the wavelength of the Raman signal will change with the different chemical structures of the test sample, which has great application value. However, due to the weak signal, it faces multiple technical obstacles in practical applications. Professor Chu Kaiqin and his team have devoted themselves to the application and research of Raman signal analysis technology, and through the research of single-particle optical tweezers automatic capture and release technology, phase-guided Raman analysis, and Raman signal algorithm analysis, Raman signal analysis technology has been applied in specific research projects, and the "label-free" biochemical detection technology has achieved breakthroughs in many aspects.Next, I would like to introduce to you a series of achievements of the "kicking hall" traditional inspection and testing method developed by Professor Chu Kaiqin's team. The content is longer and optional**, and professionals can find **address** at the end of each system chapter**. Nanoparticle multi-parameter characterization analysis system
This is a platform composed of multi-modal optical tweezers system + integrated control system + multi-modal data analysis system, which can use Raman signal analysis technology to quickly characterize the physicochemical and biochemical properties of functionalized nanoparticles and natural biological nanoparticles such as viruses, bacteria, exosomes, and sub-organelles at the "single particle" level for particles with particle sizes in the range of 100nm-500nm.
Three major features of the system:
The "single particle" characterization technology realizes the simultaneous characterization of multiple parameters of a single particle, such as size, refractive index, absolute distribution of components, and other information. It is especially suitable for the analysis of samples with high heterogeneity of single particles.
Supports automated capture and release of single particles in sample solutions.
With the support of deep learning algorithms, in the spectral analysis of complete and underdetermined chemical models, the background is automatically removed to achieve high-speed calculation.
System Parameters:
Applicable size range: 100nm-500nm;
Characterize throughput for 20-100 h;
The fastest acquisition time of Raman spectral signal <10s;
The characterization accuracy errors of size, refractive index and chemical composition are not higher than % and 5%, respectively.
Applications:
Physical, pharmaceutical, energy materials, environment, functional products and other fields.
Such as: exosome size and chemical composition analysis;
Pharmacokinetic research of liposome encapsulation technology;
Detection and analysis of drug encapsulation rate;
Characterization and analysis of individual particle parameters for other heterogeneities.
Experimental case show:
Figure 1 System test results.
We use standard nanospheres of different sizes and materials for validation, drug-loaded liposomes and exosomes of clinical plasma**. (Above) Characterization errors of particle size, refractive index, and chemical composition of the calibration system on standard particles. (middle) Characterization of the system on nanoliposomes can be further verified (bottom) Quantitative component analysis and layering analysis of nanoparticles can be performed through multimodal data. The error rates of particle size, refractive index and quantitative characterization accuracy of chemical components obtained by final calibration were 3-5% and 1., respectively5% and 48%。
Figure 1 (top) and (middle) also show that the particle size test results are in good agreement with the results of commercial instruments. As shown in Figure 1 (below), through the orthogonal information fusion of multimodal data, the system can further extract unique morphological information, such as laminarity (i.e., the number of phospholipid bilayers in liposomes), and its analysis results are often similar to the layering statistics of "standard" cryo-EM imaging.
Figure 2: Clinical characterization of plasma exosomes.
a) Characterization of particle size, refractive index and chemical composition of plasma exosomes. (b) Exosome cluster analysis based on PCA+HCA. (c) Component differences of different isotypes of exosomes. (d) Average spectra of exosomes of different isotypes in the training and validation sets.
The figures and data are all from our project team experiments, and 4 papers have been published in journals such as Analytical Chemistry, The Analyst, and BioMed Opt Express.
Citation**:1. combined morpho-chemical profiling of individual extracellular vesicles and functional nanoparticles without labels，yichuan dai, suwen bai, chuanzhen hu, kaiqin chu, bing shen*, and zachary j. smith*
anal. chem. 2020, 92, 7, 5585–5594
2. hybrid principal component analysis denoising enables rapid, label-free morpho-chemical quantification of individual nanoliposomes，yichuan dai*, yajun yu, xianli wang, ziling jiang, yulan chen, kaiqin chu, and zachary j. smith*
anal. chem. 2022, 94, 41, 14232–14241
3. multicomponent raman spectral regression using complete and incomplete models and convolutional neural networks†，derrick boateng, chuanzhen hu, yichuan dai, kaiqin chu, jun du* and zachary j. smith *
the analyst. 2022, 147(20)
4. deep learning-based size prediction for optical trapped nanoparticles and extracellular vesicles from limited bandwidth camera detection，derrick boateng, kaiqin chu, zachary j. smith, jun du,and yichuan dai
biomed opt express. 2024 jan 1; 15(1): 1–13.
Phase-Guided Raman Sampling and Analysis System
This is a new observation scheme composed of label-free phase microscopy + deep learning + Raman spectroscopy, which can realize long-term, label-free, and rapid "morphochemical" multimodal observation of organelles. The system can be used to characterize the morphological information (size and position) and contextual information (relationship with other organelles, such as lipid droplets to mitochondria, lipid droplets, and nuclei) of organelles through image acquisition and analysis, which can be used in correlation analysis.
Four major features of the system:
The detection speed is fast, and it can take even one day to obtain a signal-to-noise (SNR) organelle population spectrum within a single cell during a traditional Raman microscopy scanning process. We were able to easily acquire SNR in a controlled measurement time (approximately 10 min per cell) and each spectrum had an SNR. Dynamic tracking of submicroscopic organelles (e.g., lipid droplets) and in vivo characterization of biochemical parameters. Overcoupled high-resolution phase imaging and Raman spectroscopy enables accurate collection of spectra from single organelles within organelle clusters. For organelles with non-spatial homogeneity (e.g., phase-separated giant unilamellar vesicles), a hybrid sampling method can be used, in which the particles (in the excitation spot) are sampled at a single point and the particles are scanned in a raster over their spatial extent. However, the phase channel guidance of the system is used to "skip" the uninterested regions of the cell sample for rapid detection. 
Figure 3: System phase image and Raman spectroscopy analysis
Figure 4: System analysis function.
Applications:
It can characterize and analyze the location, quantity, environment, and chemical composition of organelles such as lipid droplets, mitochondria, and lysosomes.
Spectral changes in lipid droplets and mitochondria can be used to study the mechanism of material exchange between various organelles.
It can study the morphology and composition changes of lipid droplets in cells under drug treatment, and provide multimodal information and quantitative analysis for related drug detection.
Drug research, lipid metabolism research, and more.
Experimental case show:
Figure 5: Morphological and compositional analysis of lipid droplets.
By combining phase microscopy and Raman spectroscopy, we were able to quickly quantify the morphology and composition of lipid droplets inside fixed cells at a speed of about 600 times faster than conventional point-scanning Raman microscopy.
Figure 6: Precise quantification of lipid droplet morphology and chemical composition.
Fig.7 Continuous observation and analysis of lipid droplets3.
The figures and data are all from our project team experiments, and 1 paper has been published in analytical chemistry.
Citation**:
1.rapidintracellulardetectionandanalysisof lipiddroplets'morpho-chemical compositionby phase-guided ramansampling
haozhang,jingdefang,yichuandai,yangpan,kaiqinchu,*andzacharyj.smith*，citethis:anal.chem.2023,95,13555−13565
Organelle-specific phase contrast microscopy
The function of this system is to automatically analyze the kinetic parameters and interactions of multiple organelles in unlabeled cells, enabling label-free visualization and analysis of whole organelle dynamics and interactions.
Four major features of the system:
Mitochondria and lysosomes can be identified and tracked through a single 2D intensity image, the image acquisition is completed by the microscope, and the dynamic recognition is all completed by the algorithm, with a high degree of intelligence.
Imaging with low photon flux (approximately 40 times lower than the flux of laser-excited fluorescence irradiation) and observation and automated analysis of mitochondrial fission and fusion rates with minimal interference yields results closer to the theoretical unobserved state.
To quantitatively analyze the results, we calculated several metrics such as the Structural Similarity Index (SSIM), Multiscale SSIM (MS-SSIM), and Pearson correlation coefficients. Through the quantitative analysis of the above three indicators, we can systematically evaluate the accuracy and reliability of the proposed algorithm or model in terms of intracellular lysosomal and mitochondrial location and morphology.
We can also add a fluorescence channel to analyze the interrelationships between the four structures of cell membrane, nucleus, mitochondria, and protein with a single label (no need for multiple fluorescence channels and multiple staining).
Applications:Observe the interaction between various organelles (including mitochondria, lysosomes, endoplasm, lipid droplets, etc.).
Stress status and quantitative statistical analysis of mitochondria, lysosomes and other organelles under physical stimuli (e.g., low temperature, pH, starvation, etc.) and drug stimuli (e.g., FCCP, Erestin, etc.). Mitochondrial fusion and fusion, lysosomal movement, and detection of mitochondria-lysosomal contact sites (MLCs). Observables:Mitochondria, lysosomes, endoplasm, lipid droplets, vesicles, chromosomes, nucleus, cell membranes, etc.
Experimental case show:
Fig.8 Mitochondria interact with DRP1
Our system can see mitochondria, DRP1, and endoplasm dynamics similar to those found in conventional fluorescence imaging. (a) DRP1 is involved in mitochondrial fission; (b) DRP1 causes mitochondrial contraction without fission; (c) Directional migration of DRP1 to mitochondria; (D) DRP1 follows mitochondrial motility.
Figure 9: Cell planar analysis of mitochondrial and DRP1 egg distribution.
a) Identification of cell membranes, nuclei and mitochondria; (b) Mitochondrial to nucleus distance: the relative distance from mitochondria to nucleus; (c) The distribution of mitochondria to the nucleus and cell membrane has a consistent trend; (d) Extraction of DRP1 aggregates by rolling ball method and threshold treatment; (e) DRP1 distribution**perichondria, 30% of DRP1 co-localized with mitochondria; (e) Mitochondrial distance from DRP1.
Figure 10: Analysis of mitochondrial dynamics.
At the top of Figure A, the mitochondria are two mitochondria at 73 sec. At the bottom, the two mitochondria fuse into mitochondria at 225 sec. The initial mitochondrial image and movement over a 10-min period were presented. d f shows the release part. Not only can you see the starting position (dark blue dot table) and end position (red dot table), but you can also see the movement trajectory e and f in this time period, and the fission and fusion process is unfolded in three frames, and the two mitochondrial lines before or after fission are connected.
Figure 11: Analysis of lysosomal dynamics.
A and B are images of normally cultured and starved cells. It can be observed that in starved cells, organelles tend to gather around the nucleus more than normal cells. The locomotion trajectories of each individual lysosome can be tracked, with individual examples of normal cells exhibited in Figure 11c. We then calculated the mean displacement (MSD) of each lysosome. By fitting the MSD with the anomaly diffusion model, we obtained the anomalous diffusion factors (the results of the analysis of 10 randomly selected trajectories in 5 sec are shown in Figures 11d and e). Figure 11f shows the distribution of all lysosomal trajectories (10 cells per group). The range of motion of lysosomes in starved cells is more limited than in normal cells (1). Given that lysosomes tend to clump around the nucleus in starved cells (Figure 11B), their range of motion is expected to be more limited. In addition, we can also reflect the state of the whole cell by cell-level values (i.e., the average of all trajectories within a single cell). The results are shown in Figure 11g. It can be seen that the values of starved cells are compared to those of normal cells, which further confirms the conclusions in Figure 11f.
The figures and data are all from our project team experiments, and 1 paper was published in ACS Photonics.
Citation**:
1.label-free analysis of organelle interactions using organellespecific phase contrast microscopy (os-pcm)jingde fang, hao zhang, yang pan, zachary j.smith,* and kaiqin chu*cite this:acs photonics 2023, 10, 1093 1103 How to prevent academic fraudDM our team to learn more