Machine learning using convolutional neural networks for SERS analysis of biomarkers in medical diagnostics

SERS microscopy as a tool for comprehensive biochemical characterization in complex samples

Surface enhanced Raman scattering (SERS) spectra of biomaterials such as cells or tissues can be used to obtain biochemical information from nanoscopic volumes in these heterogeneous samples. This tutorial review discusses the factors that determine the outcome of a SERS experiment in complex bioorganic samples. They are related to the SERS process itself, the possibility to selectively probe certain regions or constituents of a sample, and the retrieval of the vibrational information in order to identify molecules and their interaction. After introducing basic aspects of SERS experiments in the context of biocompatible environments, spectroscopy in typical microscopic settings is exemplified, including the possibilities to combine SERS with other linear and non-linear microscopic tools, and to exploit approaches that improve lateral and temporal resolution. In particular the great variation of data in a SERS experiment calls for robust data analysis tools. Approaches will be introduced that have been originally developed in the field of bioinformatics for the application to omics data and that show specific potential in the analysis of SERS data. They include the use of simulated data and machine learning tools that can yield chemical information beyond achieving spectral classification.

Paper/GO/e-Au flexible SERS sensors for in situ detection of tricyclazole in orange juice and on cucumber skin at the sub-ppb level: machine learning-assisted data analysis

Despite being an excellent surface enhanced Raman scattering (SERS) active material, gold nanoparticles were difficult to be loaded onto the surface of filter paper to fabricate flexible SERS substrates. In this study, electrochemically synthesized gold nanoparticles (e-AuNPs) were deposited on graphene oxide (GO) nanosheets in solution by ultrasonication, resulting in the formation of a GO/Au hybrid material. Thanks to the support of GO, the hybrid material could adhere onto the surface of filter paper, which was immersed into a GO/Au solution for 24 h and dried naturally at room temperature. The paper-based materials were then employed as substrates for a surface enhanced Raman scattering (SERS) sensing platform to detect tricyclazole (TCZ), a widely used pesticide, resulting in better sensitivity compared to the use of paper/Au SERS sensors. With the most optimal GO content of 4%, paper/GO/Au SERS sensors could achieve a limit of detection of 1.32 × 10-10 M in standard solutions. Furthermore, the filter paper-based SERS sensors also exhibited significant advantages in sample collection in real samples. On one hand, the sensors were dipped into orange juice, allowing TCZ molecules in this real sample to be adsorbed onto their SERS active surface. On the other hand, they were pasted onto cucumber skin to collect the analytes. As a result, the paper/GO/Au SERS sensors could sense TCZ in orange juice and on cucumber skin at concentrations as low as 10-9 M (∼2 ppb). In addition, a machine learning model was designed and developed, allowing the sensing system to discriminate TCZ from nine other organic compounds and predict the presence of TCZ on cucumber skin at concentrations down to 10-9 M.

RamanFormer: A Transformer-Based Quantification Approach for Raman Mixture Components

Raman spectroscopy is a noninvasive technique to identify materials by their unique molecular vibrational fingerprints. However, distinguishing and quantifying components in mixtures present challenges due to overlapping spectra, especially when components share similar features. This study presents "RamanFormer", a transformer-based model designed to enhance the analysis of Raman spectroscopy data. By effectively managing sequential data and integrating self-attention mechanisms, RamanFormer identifies and quantifies components in chemical mixtures with high precision, achieving a mean absolute error of 1.4% and a root mean squared error of 1.6%, significantly outperforming traditional methods such as least squares, MLP, VGG11, and ResNet50. Tested extensively on binary and ternary mixtures under varying conditions, including noise levels with a signal-to-noise ratio of up to 10 dB, RamanFormer proves to be a robust tool, improving the reliability of material identification and broadening the application of Raman spectroscopy in fields, such as material science, forensics, and biomedical diagnostics.

Rapid discrimination of four Salmonella enterica serovars: A performance comparison between benchtop and handheld Raman spectrometers

Foodborne illnesses, particularly those caused by Salmonella enterica with its extensive array of over 2600 serovars, present a significant public health challenge. Therefore, prompt and precise identification of S. enterica serovars is essential for clinical relevance, which facilitates the understanding of S. enterica transmission routes and the determination of outbreak sources. Classical serotyping methods via molecular subtyping and genomic markers currently suffer from various limitations, such as labour intensiveness, time consumption, etc. Therefore, there is a pressing need to develop new diagnostic techniques. Surface-enhanced Raman spectroscopy (SERS) is a non-invasive diagnostic technique that can generate Raman spectra, based on which rapid and accurate discrimination of bacterial pathogens could be achieved. To generate SERS spectra, a Raman spectrometer is needed to detect and collect signals, which are divided into two types: the expensive benchtop spectrometer and the inexpensive handheld spectrometer. In this study, we compared the performance of two Raman spectrometers to discriminate four closely associated S. enterica serovars, that is, S. enterica subsp. enterica serovar dublin, enteritidis, typhi and typhimurium. Six machine learning algorithms were applied to analyse these SERS spectra. The support vector machine (SVM) model showed the highest accuracy for both handheld (99.97%) and benchtop (99.38%) Raman spectrometers. This study demonstrated that handheld Raman spectrometers achieved similar prediction accuracy as benchtop spectrometers when combined with machine learning models, providing an effective solution for rapid, accurate and cost-effective identification of closely associated S. enterica serovars.

Spectral encoder to extract the efficient features of Raman spectra for reliable and precise quantitative analysis

Raman spectroscopy has become a powerful analytical tool highly demanded in many applications such as microorganism sample analysis, food quality control, environmental science, and pharmaceutical analysis, owing to its non-invasiveness, simplicity, rapidity and ease of use. Among them, quantitative research using Raman spectroscopy is a crucial application field of spectral analysis. However, the entire process of quantitative modeling largely relies on the extraction of effective spectral features, particularly for measurements on complex samples or in environments with poor spectral signal quality. In this paper, we propose a method of utilizing a spectral encoder to extract effective spectral features, which can significantly enhance the reliability and precision of quantitative analysis. We built a latent encoded feature regression model; in the process of utilizing the autoencoder for reconstructing the spectrometer output, the latent feature obtained from the intermediate bottleneck layer is extracted. Then, these latent features are fed into a deep regression model for component concentration prediction. Through detailed ablation and comparative experiments, our proposed model demonstrates superior performance to common methods on single-component and multi-component mixture datasets, remarkably improving regression precision while without needing user-selected parameters and eliminating the interference of irrelevant and redundant information. Furthermore, in-depth analysis reveals that latent encoded feature possesses strong nonlinear feature representation capabilities, low computational costs, wide adaptability, and robustness against noise interference. This highlights its effectiveness in spectral regression tasks and indicates its potential in other application fields. Sufficient experimental results show that our proposed method provides a novel and effective feature extraction approach for spectral analysis, which is simple, suitable for various methods, and can meet the measurement needs of different real-world scenarios.

Machine Learning for COVID-19 Determination Using Surface-Enhanced Raman Spectroscopy

The rapid, low cost, and efficient detection of SARS-CoV-2 virus infection, especially in clinical samples, remains a major challenge. A promising solution to this problem is the combination of a spectroscopic technique: surface-enhanced Raman spectroscopy (SERS) with advanced chemometrics based on machine learning (ML) algorithms. In the present study, we conducted SERS investigations of saliva and nasopharyngeal swabs taken from a cohort of patients (saliva: 175; nasopharyngeal swabs: 114). Obtained SERS spectra were analyzed using a range of classifiers in which random forest (RF) achieved the best results, e.g., for saliva, the precision and recall equals 94.0% and 88.9%, respectively. The results demonstrate that even with a relatively small number of clinical samples, the combination of SERS and shallow machine learning can be used to identify SARS-CoV-2 virus in clinical practice.

Surface-Enhanced Raman Spectroscopy-Based Detection of Micro-RNA Biomarkers for Biomedical Diagnosis Using a Comparative Study of Interpretable Machine Learning Algorithms

Surface-enhanced Raman spectroscopy (SERS) has wide diagnostic applications due to narrow spectral features that allow multiplex analysis. We have previously developed a multiplexed, SERS-based nanosensor for micro-RNA (miRNA) detection called the inverse molecular sentinel (iMS). Machine learning (ML) algorithms have been increasingly adopted for spectral analysis due to their ability to discover underlying patterns and relationships within large and complex data sets. However, the high dimensionality of SERS data poses a challenge for traditional ML techniques, which can be prone to overfitting and poor generalization. Non-negative matrix factorization (NMF) reduces the dimensionality of SERS data while preserving information content. In this paper, we compared the performance of ML methods including convolutional neural network (CNN), support vector regression, and extreme gradient boosting combined with and without NMF for spectral unmixing of four-way multiplexed SERS spectra from iMS assays used for miRNA detection. CNN achieved high accuracy in spectral unmixing. Incorporating NMF before CNN drastically decreased memory and training demands without sacrificing model performance on SERS spectral unmixing. Additionally, models were interpreted using gradient class activation maps and partial dependency plots to understand predictions. These models were used to analyze clinical SERS data from single-plexed iMS in RNA extracted from 17 endoscopic tissue biopsies. CNN and CNN-NMF, trained on multiplexed data, performed most accurately with RMSElabel = 0.101 and 9.68 × 10-2, respectively. We demonstrated that CNN-based ML shows great promise in spectral unmixing of multiplexed SERS spectra, and the effect of dimensionality reduction on performance and training speed.

Comprehensive modeling of cell culture profile using Raman spectroscopy and machine learning

Chinese hamster ovary (CHO) cells are widely utilized in the production of antibody drugs. To ensure the production of large quantities of antibodies that meet the required specifications, it is crucial to monitor and control the levels of metabolites comprehensively during CHO cell culture. In recent years, continuous analysis methods employing on-line/in-line techniques using Raman spectroscopy have attracted attention. While these analytical methods can nondestructively monitor culture data, constructing a highly accurate measurement model for numerous components is time-consuming, making it challenging to implement in the rapid research and development of pharmaceutical manufacturing processes. In this study, we developed a comprehensive, simple, and automated method for constructing a Raman model of various components measured by LC-MS and other techniques using machine learning with Python. Preprocessing and spectral-range optimization of data for model construction (partial least square (PLS) regression) were automated and accelerated using Bayes optimization. Subsequently, models were constructed for each component using various model construction techniques, including linear regression, ridge regression, XGBoost, and neural network. This enabled the model accuracy to be improved compared with PLS regression. This automated approach allows continuous monitoring of various parameters for over 100 components, facilitating process optimization and process monitoring of CHO cells.

Multiplex SERS detection of polycyclic aromatic hydrocarbon (PAH) pollutants in water samples using gold nanostars and machine learning analysis

Polycyclic aromatic hydrocarbons (PAHs) have attracted a lot of environmental concern because of their carcinogenic and mutagenic properties, and the fact they can easily contaminate natural resources such as drinking water and river water. This study presents a simple and sensitive point-of-care SERS detection of PAHs combined with machine learning algorithms to predict the PAH content more precisely and accurately in real-life samples such as drinking water and river water. We first synthesized multibranched sharp-spiked surfactant-free gold nanostars (GNSs) that can generate strong surface-enhanced Raman scattering (SERS) signals, which were further coated with cetyltrimethylammonium bromide (CTAB) for long-term stability of the GNSs as well as to trap PAHs. We utilized CTAB-capped GNSs for solution-based 'mix and detect' SERS sensing of various PAHs including pyrene (PY), nitro-pyrene (NP), anthracene (ANT), benzo[a]pyrene (BAP), and triphenylene (TP) spiked in drinking water and river water using a portable Raman module. Very low limits of detection (LOD) were achieved in the nanomolar range for the PAHs investigated. More importantly, the detected SERS signal was reproducible for over 90 days after synthesis. Furthermore, we analyzed the SERS data using artificial intelligence (AI) with machine learning algorithms based on the convolutional neural network (CNN) model in order to discriminate the PAHs in samples more precisely and accurately. Using a CNN classification model, we achieved a high prediction accuracy of 90% in the nanomolar detection range and an f1 score (harmonic mean of precision and recall) of 94%, and using a CNN regression model, achieved an RMSEconc = 1.07 × 10-1 μM. Overall, our SERS platform can be effectively and efficiently used for the accurate detection of PAHs in real-life samples, thus opening up a new, sensitive, selective, and practical approach for point-of-need SERS diagnosis of small molecules in complex practical environments.

Dense Convolutional Neural Network for Identification of Raman Spectra

The rapid development of cloud computing and deep learning makes the intelligent modes of applications widespread in various fields. The identification of Raman spectra can be realized in the cloud, due to its powerful computing, abundant spectral databases and advanced algorithms. Thus, it can reduce the dependence on the performance of the terminal instruments. However, the complexity of the detection environment can cause great interferences, which might significantly decrease the identification accuracies of algorithms. In this paper, a deep learning algorithm based on the Dense network has been proposed to satisfy the realization of this vision. The proposed Dense convolutional neural network has a very deep structure of over 40 layers and plenty of parameters to adjust the weight of different wavebands. In the kernel Dense blocks part of the network, it has a feed-forward fashion of connection for each layer to every other layer. It can alleviate the gradient vanishing or explosion problems, strengthen feature propagations, encourage feature reuses and enhance training efficiency. The network's special architecture mitigates noise interferences and ensures precise identification. The Dense network shows more accuracy and robustness compared to other CNN-based algorithms. We set up a database of 1600 Raman spectra consisting of 32 different types of liquid chemicals. They are detected using different postures as examples of interfered Raman spectra. In the 50 repeated training and testing sets, the Dense network can achieve a weighted accuracy of 99.99%. We have also tested the RRUFF database and the Dense network has a good performance. The proposed approach advances cloud-enabled Raman spectra identification, offering improved accuracy and adaptability for diverse identification tasks.

Unraveling surface-enhanced Raman spectroscopy results through chemometrics and machine learning: principles, progress, and trends

Surface-enhanced Raman spectroscopy (SERS) has gained increasing attention because it provides rich chemical information and high sensitivity, being applicable in many scientific fields including medical diagnosis, forensic analysis, food control, and microbiology. Although SERS is often limited by the lack of selectivity in the analysis of samples with complex matrices, the use of multivariate statistics and mathematical tools has been demonstrated to be an efficient strategy to circumvent this issue. Importantly, since the rapid development of artificial intelligence has been promoting the implementation of a wide variety of advanced multivariate methods in SERS, a discussion about the extent of their synergy and possible standardization becomes necessary. This critical review comprises the principles, advantages, and limitations of coupling SERS with chemometrics and machine learning for both qualitative and quantitative analytical applications. Recent advances and trends in combining SERS with uncommonly used but powerful data analysis tools are also discussed. Finally, a section on benchmarking and tips for selecting the suitable chemometric/machine learning method is included. We believe this will help to move SERS from an alternative detection strategy to a general analytical technique for real-life applications.

Deep Learning-Assisted Surface-Enhanced Raman Scattering for Rapid Bacterial Identification

Bloodstream infection (BSI) is characterized by the presence of viable microorganisms in the bloodstream and may induce systemic immune responses. Early and appropriate antibiotic usage is crucial to effectively treating BSI. However, conventional culture-based microbiological diagnostics are time-consuming and cannot provide timely bacterial identification for subsequent antimicrobial susceptibility test (AST) and clinical decision-making. To address this issue, modern microbiological diagnostics have been developed, such as surface-enhanced Raman scattering (SERS), which is a sensitive, label-free, and quick bacterial detection method measuring specific bacterial metabolites. In this study, we aim to integrate a new deep learning (DL) method, Vision Transformer (ViT), with bacterial SERS spectral analysis to build the SERS-DL model for rapid identification of Gram type, species, and resistant strains. To demonstrate the feasibility of our approach, we used 11,774 SERS spectra obtained directly from eight common bacterial species in clinical blood samples without artificial introduction as the training dataset for the SERS-DL model. Our results showed that ViT achieved excellent identification accuracy of 99.30% for Gram type and 97.56% for species. Moreover, we employed transfer learning by using the Gram-positive species identifier as a pre-trained model to perform the antibiotic-resistant strain task. The identification accuracy of methicillin-resistant and -susceptible Staphylococcus aureus (MRSA and MSSA) can reach 98.5% with only 200-dataset requirement. In summary, our SERS-DL model has great potential to provide a quick clinical reference to determine the bacterial Gram type, species, and even resistant strains, which can guide early antibiotic usage in BSI.

Design, Rationalization, and Automation of a Catalytic Sensing Mechanism for Homogeneous SERS Biosensors

The current pandemic has shown that we need sensitive and deployable diagnostic technologies. Surface-enhanced Raman scattering (SERS) sensors can be an ideal solution for developing such advanced point-of-need (PON) diagnostic tests. Homogeneous (reagentless) SERS sensors work by directly responding to the target without any processing step, making them capable for simple one-pot assays, but their limitation is the achievable sensitivity, insufficient compared to what is needed for sensing of viral biomarkers. Noncovalent DNA catalysis mechanisms have been recently exploited for catalytic amplification in SERS assays. These advances used catalytic hairpin assembly (CHA) and other DNA self-assembly processes to develop sensing mechanisms with improved sensitivities. However, these mechanisms have not been used in OFF-to-ON homogeneous sensors, and they often target the same biomarker, likely due to the complexity of the mechanism design. There is still a strong need for a catalytic SERS sensor with a homogeneous mechanism and a rationalization of the catalytic sensing mechanism to translate this sensing strategy to different targets and applications. We developed and investigated a homogeneous SERS sensing mechanism that uses catalytic amplification based on DNA self-assembly. We systematically investigated the role of three domains in the fuel strand (internal loop, stem, and toehold), which drives the catalytic mechanism. The thermodynamic parameters determined in our studies were used to build an algorithm for automated design of catalytic sensors that we validated on target sequences associated with malaria and SARS-CoV-2 strains. With our mechanism, we were able to achieve an amplification level of 20-fold for conventional DNA and of 36-fold using locked nucleic acids (LNAs), with corresponding improvements observed in the sensor limit of detection (LOD). We also show a single-base sequence specificity for a sensor targeting a sequence associated with the omicron variant, tested against a delta variant target. This work on catalytic amplification of homogeneous SERS sensors has the potential to enable the use of this sensing modality in new applications, such as infectious disease surveillance, by improving the LOD while conserving the sensor's homogeneous character.

Advances in surface-enhanced Raman spectroscopy technology for detection of foodborne pathogens

Rapid control and prevention of diseases caused by foodborne pathogens is one of the existing food safety regulatory issues faced by various countries and has received wide attention from all sectors of society. The development of rapid and reliable detection methods for foodborne pathogens remains a hot research area for food safety and public health because of the limitations of complex steps, time-consuming, low sensitivity, or poor selectivity of commonly used methods. Surface-enhanced Raman spectroscopy (SERS), as a novel spectroscopic technique, has the advantages of high sensitivity, selectivity, rapid and nondestructive detection and has exhibited broad application prospects in the determination of pathogenic bacteria. In this study, the enhancement mechanisms of SERS are briefly introduced, then the characteristics and properties of liquid-phase, rigid solid-phase, and flexible solid-phase are categorized. Furthermore, a comprehensive review of the advances in label-free or label-based SERS strategies and SERS-compatible techniques for the detection of foodborne pathogens is provided, and the advantages and disadvantages of these methods are reviewed. Finally, the current challenges of SERS technology applied in practical applications are listed, and the possible development trends of SERS in the field of foodborne pathogens detection in the future are discussed.

Application of serum SERS technology based on thermally annealed silver nanoparticle composite substrate in breast cancer

Liquid biopsy is currently a non-destructive and convenient method of cancer screening, due to human blood containing a variety of cancer-related biomolecules. Therefore, the development of an accurate and rapid breast cancer screening technique combined with breast cancer serum is crucial for the treatment and prognosis of breast cancer patients. In this study, the surface enhanced Raman spectroscopy (SERS) technique is used to enhance the Raman spectroscopy (RS) signal of serum based on a high sensitivity thermally annealed silver nanoparticle/porous silicon bragg mirror (AgNPs/PSB) composite substrate. Compared with RS, SERS reflects more and stronger spectral peak information, which is beneficial to discover new biomarkers of breast cancer. At the same time, to further explore the diagnostic ability of SERS technology for breast cancer. In this study, the raw spectral data are processed by baseline correction, polynomial smoothing, and normalization. Then, the relevant feature information of SERS and RS is extracted by principal component analysis (PCA), and five classification models are established to compare the diagnostic performance of SERS and RS models respectively. The experimental results show that the breast cancer diagnosis model based on the improved SERS substrate combined with the machine learning algorithm can be used to distinguish breast cancer patients from controls. The accuracy, sensitivity, specificity and AUC values of the SVM model are 100%, 100%, 100% and 100%, respectively, as well as the training time of 4ms. The above experimental results show that the SERS technology based on AgNPs/PSB composite substrate, combined with machine learning methods, has great potential in the rapid and accurate identification of breast cancer patients.