Article

Rapid Identification of Urinary Tract Infection Bacteria Using Hyperspectral, Whole Organism Fingerprinting and Artificial Neural Networks

Authors:
  • Aberystwyth University, Aberystwyth, Wales, UK
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Three rapid spectroscopic approaches for whole-organism fingerprinting-pyrolysis mass spectrometry (PyMS), Fourier transform infra-red spectroscopy (FT-IR) and dispersive Raman microscopy--were used to analyse a group of 59 clinical bacterial isolates associated with urinary tract infection. Direct visual analysis of these spectra was not possible, highlighting the need to use methods to reduce the dimensionality of these hyperspectral data. The unsupervised methods of discriminant function and hierarchical cluster analyses were employed to group these organisms based on their spectral fingerprints, but none produced wholly satisfactory groupings which were characteristic for each of the five bacterial types. In contrast, for PyMS and FT-IR, the artificial neural network (ANN) approaches exploiting multi-layer perceptrons or radial basis functions could be trained with representative spectra of the five bacterial groups so that isolates from clinical bacteriuria in an independent unseen test set could be correctly identified. Comparable ANNs trained with Raman spectra correctly identified some 80% of the same test set. PyMS and FT-IR have often been exploited within microbial systematics, but these are believed to be the first published data showing the ability of dispersive Raman microscopy to discriminate clinically significant intact bacterial species. These results demonstrate that modern analytical spectroscopies of high intrinsic dimensionality can provide rapid accurate microbial characterization techniques, but only when combined with appropriate chemometrics.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... SERS mediated uropathogen identification has been achieved with accuracies of up to 95.8% (e.g. 18,21,34,[38][39][40][41][42][43][44] ). However, circumventing bacterial amplification has required complex processes such as immunocapture, dielectrophoresis and use of optical tweezers to capture or aggregate pathogens, which are technically challenging or require expensive equipment 39,45 . ...
... Moreover, as vacuum filtration allows for rapid separation of uropathogens from urine, direct classification was achieved in urine samples rather than in suspensions or precultured samples. Additionally, a handheld Raman spectrometer was used in this study rather than the Raman microscopes commonly used in other SERS work 18,34,[38][39][40]43 . While Raman microscopes provide high-resolution spectra, the high costs, large footprints and requirements for technical expertise preclude their incorporation into point-of-care devices. ...
Article
Full-text available
Urinary tract infection is one of the most common bacterial infections leading to increased morbidity, mortality and societal costs. Current diagnostics exacerbate this problem due to an inability to provide timely pathogen identification. Surface enhanced Raman spectroscopy (SERS) has the potential to overcome these issues by providing immediate bacterial classification. To date, achieving accurate classification has required technically complicated processes to capture pathogens, which has precluded the integration of SERS into rapid diagnostics. This work demonstrates that gold-coated membrane filters capture and aggregate bacteria, separating them from urine, while also providing Raman signal enhancement. An optimal gold coating thickness of 50 nm was demonstrated, and the diagnostic performance of the SERS-active filters was assessed using phantom urine infection samples at clinically relevant concentrations (10⁵ CFU/ml). Infected and uninfected (control) samples were identified with an accuracy of 91.1%. Amongst infected samples only, classification of three bacteria (Escherichia coli, Enterococcus faecalis, Klebsiella pneumoniae) was achieved at a rate of 91.6%.
... The typical structure of a QSAR type of problem is provided in Figure 2A, where a series of molecules represented as SMILES strings [27] are encoded as molecular fingerprints and used to learn a nonlinear mapping to produce an output in the form of a classification or regression estimation. The architecture of this is implicitly in the form of a multilayer perceptron (a classical neural network [28][29][30][31]), in which weights are modified ("trained") to provide a mapping from input SMILES to a numerical output. Our fundamental problem stems from the fact that these types of encoding are one-way: the SMILES string can generate the molecular fingerprint but the molecular fingerprint cannot generate the SMILES. ...
... Molecules 2020, 25, x; doi: FOR PEER REVIEW www.mdpi.com/journal/molecules [28][29][30][31]), in which weights are modified ("trained") to provide a mapping from input SMILES to a numerical output. Our fundamental problem stems from the fact that these types of encoding are one-way: the SMILES string can generate the molecular fingerprint but the molecular fingerprint cannot generate the SMILES. ...
Article
Full-text available
Molecular similarity is an elusive but core “unsupervised” cheminformatics concept, yet different “fingerprint” encodings of molecular structures return very different similarity values, even when using the same similarity metric. Each encoding may be of value when applied to other problems with objective or target functions, implying that a priori none are “better” than the others, nor than encoding-free metrics such as maximum common substructure (MCSS). We here introduce a novel approach to molecular similarity, in the form of a variational autoencoder (VAE). This learns the joint distribution p(z|x) where z is a latent vector and x are the (same) input/output data. It takes the form of a “bowtie”-shaped artificial neural network. In the middle is a “bottleneck layer” or latent vector in which inputs are transformed into, and represented as, a vector of numbers (encoding), with a reverse process (decoding) seeking to return the SMILES string that was the input. We train a VAE on over six million druglike molecules and natural products (including over one million in the final holdout set). The VAE vector distances provide a rapid and novel metric for molecular similarity that is both easily and rapidly calculated. We describe the method and its application to a typical similarity problem in cheminformatics.
... Thus, it was identified in the 1990s that machine learning methods, could efficiently discriminate between spectra of different molecules [1] and be more effective than linear regression methods for data analysis in Raman spectroscopy [2]. Hence, machine learning algorithms were then applied in fields such as food analysis [3], bacteria identification [4], diagnostic application [5], and material analysis [6]. ...
... Otherwise, overlapping of the results might occur for spectra that are not as well defined. If correctly labelled data is available, supervised learning is likely a better choice as the program will know to look for differences in spectra with different labels rather than blindly searching all the data for similarities and potentially finding incoherent differences instead [4]. ...
Article
Machine learning is shaping up our lives in many ways. In analytical sciences, machine learning provides an unprecedented opportunity to extract information from complex or big datasets in chromatography, mass spectrometry, NMR, and spectroscopy, among others. This is especially the case in Raman and surface-enhanced Raman scattering (SERS) techniques where vibrational spectra of complex chemical mixtures are acquired as large datasets for the analysis or imaging of chemical systems. The classical linear methods of processing the information no longer suffice and thus machine learning methods for extracting the chemical information from Raman and SERS experiments have been implemented recently. In this review, we will provide a brief overview of the most common machine learning techniques employed in Raman, a guideline for new users to implement machine learning in their data analysis process, and an overview of modern applications of machine learning in Raman and SERS.
... Previous research has shown that the Raman spectroscopic technique provides a good analysis of bacterial cells (He et al. 2020;Kim et al. 2020;Akanny et al. 2021;Ciloglu et al. 2021;Wang et al. 2021). This method has proven effective for the rapid identification of bacteria (Goodacre et al. 1998(Goodacre et al. , 2000Naumann 2001;Petrich 2001;Maquelin et al. 2003;Petry et al. 2003). ...
Article
Full-text available
The decline in the performance of spiral-wound reverse osmosis (SWRO) membranes is frequently due to biofouling. This study focus on qualitative and quantitative diagnosis of SWRO membrane biofouling. Bacterial counts on the different surfaces of the fouled membranes were carried out. Surface enhanced Raman spectroscopy (SERS) was performed to highlight clogging materials as well as their natures and identity. The topography of the fouled membranes and the structures of biofilms were visualized by fluorescence microscopy (FM) and scanning electron microscopy (SEM). The results indicated the presence of bacteria in the different SWRO membrane areas. Those strongly adhered were significantly higher than those weakly. It varied between 26 × 105 and 262 × 105 CFU m-2. However, SERS mapping showed different fouling levels and the thickness of the fouling layer was 5 µm. Microscopic imaging revealed biotic and abiotic deposits. These data can together allow better management of the seawater desalination process.
... Precision/recall curves for (a) the logistic regression, and (b) the 2-layer FCN models.DiscussionClassical machine learning algorithms, such as linear models, support vector machines 24 and random forests, 25 are often used for analyzing traditional MALDI-TOF MS spectra 9,26-29 , and they are also often deployed on several single-cell technologies, such as flow cytometry 30,31 and Raman spectroscopy.[32][33][34] More recently, deep learning models have been proposed for the identification of bacterial species based on Raman data.21,35 ...
Article
In diagnostics of infectious diseases, MALDI-TOF mass spectrometry (MALDI-TOF MS) can be applied for the identification of pathogenic microorganisms. However, to achieve a trustworthy identification from MALDI-TOF MS data, a significant amount of biomass should be considered. The bacterial load that potentially occurs in a sample is therefore routinely amplified by culturing, which is a time-consuming procedure. In this paper we show that culturing can be avoided by conducting MALDI-TOF MS on individual bacterial cells. This results in a more rapid identification of species with an acceptable accuracy. We propose a deep learning architecture to analyze the data and compare its performance with traditional supervised machine learning algorithms. We illustrate our workflow on a large dataset that contains bacterial species related to urinary tract infections. Overall we obtain accuracies up to 85\% in discriminating five different species.
... The Raman signal can be enhanced with metals (known as surfaced-enhanced Raman spectroscopy or SERS), which reduces the acquisition time to 1-3 s per cell (24). The resulting spectrum contains biochemical information of the molecules that are present in the cellfor example, lipids, carbohydrates, nucleic acids, and proteins-and can be used to classify bacteria according to phylogeny (25). This information can be quantitative if an internal standard for the molecule(s) of interest is made. ...
Article
Investigating phenotypic heterogeneity can help to better understand and manage microbial communities. However, characterizing phenotypic heterogeneity remains a challenge, as there is no standardized analysis framework. Several optical tools are available, such as flow cytometry and Raman spectroscopy, which describe optical properties of the individual cell. In this work, we compare Raman spectroscopy and flow cytometry to study phenotypic heterogeneity in bacterial populations. The growth stages of three replicate Escherichia coli populations were characterized using both technologies. Our findings show that flow cytometry detects and quantifies shifts in phenotypic heterogeneity at the population level due to its high‐throughput nature. Raman spectroscopy, on the other hand, offers a much higher resolution at the single‐cell level (i.e., more biochemical information is recorded). Therefore, it can identify distinct phenotypic populations when coupled with analyses tailored toward single‐cell data. In addition, it provides information about biomolecules that are present, which can be linked to cell functionality. We propose a computational workflow to distinguish between bacterial phenotypic populations using Raman spectroscopy and validated this approach with an external data set. We recommend using flow cytometry to quantify phenotypic heterogeneity at the population level, and Raman spectroscopy to perform a more in‐depth analysis of heterogeneity at the single‐cell level. © 2019 International Society for Advancement of Cytometry
... The authors reported different recognition accuracies, which vary between 94% and 98%. Goodacre et al. [5], used different spectrometry and artificial neural networks (ANN) approaches to identify bacteria types. They have reached the highest identification accuracy of 80%. ...
Preprint
Full-text available
Automated recognition and classification of bacteria species from microscopic images have significant importance in clinical microbiology. Bacteria classification is usually carried out manually by biologists using different shapes and morphologic characteristics of bacteria species. The manual taxonomy of bacteria types from microscopy images is time-consuming and a challenging task for even experienced biologists. In this study, an automated deep learning based classification approach has been proposed to classify bacterial images into different categories. The ResNet-50 pre-trained CNN architecture has been used to classify digital bacteria images into 33 categories. The transfer learning technique was employed to accelerate the training process of the network and improve the classification performance of the network. The proposed method achieved an average classification accuracy of 99.2%. The experimental results demonstrate that the proposed technique surpasses state-of-the-art methods in the literature and can be used for any type of bacteria classification tasks.
... The resulting spectrum contains biochemical information of the molecules that are present in bacteria -e.g. lipids, carbohydrates, nucleic acids and proteins -and can be used to classify bacteria according to their population (Goodacre et al., 1998;Huang et al., 2010;Strola et al., 2014). Raman spectroscopy can be used for the monitoring of compounds present in the supernatant such as, glucose, protein production or others (Lee et al., 2004) -as well as Raman reactive compounds present in the bacteria, such as chlorophylls, carotenoids and other pigments (Jehlička et al., 2014). ...
... The authors reported different recognition accuracies, which vary between 94% and 98%. Goodacre et al. [5], used different spectrometry and artificial neural networks (ANN) approaches to identify bacteria types. They have reached the highest identification accuracy of 80%. ...
... Although such MLPs could indeed be used for virtual screening (and many other purposes), they were very slow to train (radial basis function networks [44,45] (as in Figure 3B) were far quicker [46,47]), and it proved impossible to train large nets with even modest (>2) numbers of hidden layers. It is widely considered that this was simply due to the fact that the gradient fed back during the backpropagation step was increasingly small as the number of weights increased (the 'vanishing gradient' problem). ...
Article
Full-text available
The number of ‘small’ molecules that may be of interest to chemical biologists — chemical space — is enormous, but the fraction that have ever been made is tiny. Most strategies are discriminative, i.e. have involved ‘forward’ problems (have molecule, establish properties). However, we normally wish to solve the much harder generative or inverse problem (describe desired properties, find molecule). ‘Deep’ (machine) learning based on large-scale neural networks underpins technologies such as computer vision, natural language processing, driverless cars, and world-leading performance in games such as Go; it can also be applied to the solution of inverse problems in chemical biology. In particular, recent developments in deep learning admit the in silico generation of candidate molecular structures and the prediction of their properties, thereby allowing one to navigate (bio)chemical space intelligently. These methods are revolutionary but require an understanding of both (bio)chemistry and computer science to be exploited to best advantage. We give a high-level (non-mathematical) background to the deep learning revolution, and set out the crucial issue for chemical biology and informatics as a two-way mapping from the discrete nature of individual molecules to the continuous but high-dimensional latent representation that may best reflect chemical space. A variety of architectures can do this; we focus on a particular type known as variational autoencoders. We then provide some examples of recent successes of these kinds of approach, and a look towards the future.
... The technology of clustering has many applications in gene data expression [26]- [28] and gene sequence [29]. The research on both proteomics and metabolomics are developed [30], as well as on the context of protein comparison and structure prediction [31], [32]. However, the studies in this kind are usually rely on the class symbol of the clusters. ...
Article
Full-text available
Many areas of exploratory data analysis need to deal with high-dimensional data sets. Some real life data like human gene, have an inherent structure of hierarchy, which embeds multi-layer feature groups. In this paper, we propose an algorithm to search for the number of feature groups in high-dimensional data by sequential minimax method and detect the hierarchical structure of high-dimensional data. Several proper numbers of feature grouping can be discovered. The feature grouping and group weights are investigated for each group number. After the comparison of feature groupings, the multi-layer structure of feature groups are detected. The latent feature group learning (LFGL) algorithm is proposed to evaluate the effectiveness of the number of feature groups and provide a method of subspace clustering. In the experiments on several gene datasets, the proposed algorithm outstands several representative algorithms.
... For example, Raman spectroscopy is a label-free method that utilizes the light scattering phenomena to determine the unique chemical fingerprints of any molecule by probing individual chemical bond vibrations [50]. According to the literature, the Raman spectrometer has been utilized extensively for identifying single bacteria cells or bacterial colonies based on the acquired Raman spectra [51][52][53][54]. Recently, several researchers have also exploited Raman spectroscopy for identification and detection of pathogens in ear infection cases. ...
Article
Full-text available
Ear infection is one of the most commonly occurring inflammation diseases in the world, especially for children. Almost every child encounters at least one episode of ear infection before he/she reaches the age of seven. The typical treatment currently followed by physicians is visual inspection and antibiotic prescription. In most cases, a lack of improper treatment results in severe bacterial infection. Therefore, it is necessary to design and explore advanced practices for effective diagnosis. In this review paper, we present the various types of ear infection and the related pathogens responsible for middle ear infection. We outline the conventional techniques along with clinical trials using those techniques to detect ear infections. Further, we highlight the need for emerging techniques to reduce ear infection complications. Finally, we emphasize the utility of Raman spectroscopy as a prospective non-invasive technique for the identification of middle ear infection.
... Applications of FT-IR spectroscopic technique are broad in scope, and this was accommodated in the 1990s within the biosciences, as the technique was established to be mostly useful for the discrimination of axenically cultured bacteria (Goodacre et al., 1998), and work on food-related bacteria continues (Lu et al., 2011). The work undertaken during the 90s spawned a considerable amount of interest which led to other areas of research, such as metabolomics, more specifically referred to as metabolic fingerprinting. ...
... : bioRxiv preprint employed for various medical diagnostic applications [7], cellular phenotyping [8], or delineation of diseased tissue in cancer treatment [9,10]. It is also extensively used in more fundamental research where it can be used to study specific biological processes, including cellular death [11,12,13], cellular response [14], infection detection [15] and pathogen identification [16,17] On the other hand, Raman spectroscopy has a relatively slow recording rate, which is imposed by the long exposure times required to retrieve with reasonable signal-to-noise ratio the low intensity signals emitted by biomolecules. This implies that it is often challenging to measure large numbers of samples, which are often required in biology to derive relevant findings. ...
Preprint
Full-text available
Raman spectroscopy has the ability to retrieve molecular information from live biological samples non-invasively through optical means. Coupled with machine learning, it is possible to use the large amount of information contained in a Raman spectrum to create models that can predict the state of new samples based on statistical analysis from previous measurements. Furthermore, in case of linear models, the separation coefficients can be used to interpret which bands are contributing to the discrimination between experimental conditions, which correspond here to single-cell measurements of macrophages under in vitro immune stimulation. We here evaluate a typical linear method using discriminant analysis and PCA, and compare it to regularized logistic regression (Lasso). We find that the use of PCA is not beneficial to the classification performance. Furthermore, the Lasso approach yields sparse separation vectors, since it suppresses spectral coefficients which do not improve classification, making interpretation easier. To further evaluate the approach, we apply the Lasso technique to a well-defined case where protein synthesis is inhibited, and show that the separating features are consistent with RNA accumulation and protein levels depletion. Surprisingly, when Raman features are selected purely in terms of their classification power (Lasso), the selected coefficients are contained in side bands, while typical strong Raman peaks are not present in the discrimination vector. We propose that this occurs because large Raman bands are representative of a wide variety of cellular molecules and are therefore less suited for accurate classification.
... The typical structure of a QSAR type of problem is given in Fig 2A, where a series of molecules represented as SMILES strings [27] are encoded as molecular fingerprints and used to learn a nonlinear mapping to produce an output in the form of a classification or regression estimation. The architecture of this is implicitly in the form of a multilayer perceptron (a classical neural network [28][29][30][31]), in which weights are modified ('trained') to provide a mapping from input SMILES to a numerical output. Our fundamental problem stems from the fact that these types of encoding are one-way: The SMILES string can generate the molecular fingerprint but the molecular fingerprint cannot generate the SMILES. ...
Preprint
Full-text available
Molecular similarity is an elusive but core unsupervised cheminformatics concept, yet different fingerprint encodings of molecular structures return very different similarity values even when using the same similarity metric. Each encoding may be of value when applied to other problems with objective or target functions, implying that a priori none is better than the others, nor than encoding-free metrics such as maximum common substructure (MCSS). We here introduce a novel approach to molecular similarity, in the form of a variational autoencoder (VAE). This learns the joint distribution p(z|x) where z is a latent vector and x are the (same) input/output data. It takes the form of a bowtie-shaped artificial neural network. In the middle is a bottleneck layer or latent vector in which inputs are transformed into, and represented as, a vector of numbers (encoding), with a reverse process (decoding) seeking to return the SMILES string that was the input. We train a VAE on over 6 million druglike molecules and natural products (including over one million in the final holdout set). The VAE vector distances provide a rapid and novel metric for molecular similarity that is both easily and rapidly calculated. We describe the method and its application to a typical similarity problem in cheminformatics.
... For each sample, reading was taken three times. The FTIR spectra were processed according to the method of Goodacre et al. (1998) with some modifications. All data were baseline corrected by using Origin Pro Software 2016. ...
Article
Full-text available
Abstract The presence of calcium oxalate is the major obstacle in flour processing of Amorphophallus muelleri tuber since the calcium oxalate can induce skin irritation and is harmful to kidneys. The development of a rapid analytical method to detect calcium oxalate in Amorphophallus flour is required. This research was intended to evaluate the use of some analytical methods (FTIR, SEM-EDS, XRD, XRF, and titration methods) in calcium oxalate detection in Amorphophallus muelleri flour prepared from different treatments (soaking in water (W), solution of sodium bisulfite 1000 ppm (B), solution of sodium chloride salt 3% (S), solution of sodium bisulfite 1000 ppm and sodium chloride salt 3% (BS)). Results showed that the presence of oxalate in Amorphophallus flour can be detected in the FTIR spectra from the C=O group at a wavenumber of 1610 cm-1. SEM images confirmed that calcium oxalate in Amorphophallus flour existed as raphide crystals in which their quantity can be estimated by the EDS feature of SEM. The presence of calcium oxalate crystals in Amorphophallus flour can be differentiated from other salts present in the flour by XRD. XRF can be used as a rapid analytical tool to detect the presence of calcium oxalate in Amorphophallus flour. The potassium permanganate titration technique can be used as a reference method for other rapid analytical methods in detecting calcium oxalate in Amorphophallus flour.
... Deep learning models with various architectures were applied with great success in image processing [13][14][15][16] as well as speech [17,18] recognition. Not only this, but DL methods can be applied to a variety of fields; from medicine [19][20][21] and food safety [22] to particle physics [23]. It was later found that DL methods can effectively distinguish the spectra of different molecules [24] and do so more effectively than linear regression methods for data analysis in Raman spectroscopy [25]. ...
Preprint
Full-text available
Raman spectroscopy in combination with machine learning has significant promise for applications in clinical settings as a rapid, sensitive, and label-free identification method. These approaches perform well in classifying data that contains classes that occur during the training phase. However, in practice, there are always substances whose spectra have not yet been taken or are not yet known and when the input data are far from the training set and include new classes that were not seen at the training stage, a significant number of false positives are recorded which limits the clinical relevance of these algorithms. Here we show that these obstacles can be overcome by implementing recently introduced Entropic Open Set and Objectosphere loss functions. To demonstrate the efficiency of this approach, we compiled a database of Raman spectra of 40 chemical classes separating them into 20 biologically relevant classes comprised of amino acids, 10 irrelevant classes comprised of bio-related chemicals, and 10 classes that the Neural Network has not seen before, comprised of a variety of other chemicals. We show that this approach enables the network to effectively identify the unknown classes while preserving high accuracy on the known ones, dramatically reducing the number of false positives while preserving high accuracy on the known classes, which will allow this technique to bridge the gap between laboratory experiments and clinical applications.
... Comprehensive measurements of molecular com position of samples and identification of unknown bio chemical compounds may eventually become feasible. For example, there have been a few case studies in which machine learning platforms have been trained using Raman spectra of diverse types of pathogenic bacteria and were then able to successfully assign newly meas ured data to closely related groups 106,211,212 . By contrast, Raman microspectroscopy has rarely been applied in microbiology for naive identification or measurement of unknown biochemical compounds. ...
Article
Full-text available
Raman microspectroscopy offers microbiologists a rapid and non-destructive technique to assess the chemical composition of individual live microorganisms in near real time. In this Primer, we outline the methodology and potential for its application to microbiology. We describe the technical aspects of Raman analyses and practical approaches to apply this method to microbiological questions. We discuss recent and potential future applications to determine the composition and distribution of microbial metabolites down to subcellular scale; to investigate the host–microorganism, cell–cell and cell–environment molecular exchanges that underlie the structure of microbial ecosystems from the ocean to the human gut microbiomes; and to interrogate the microbial diversity of functional roles in environmental and industrial processes — key themes in modern microbiology. We describe the current technical limitations of Raman microspectroscopy for investigation of microorganisms and approaches to minimize or address them. Recent technological innovations in Raman microspectroscopy will further reinforce the power and capacity of this method for broader adoptions in microbiology, allowing microbiologists to deepen their understanding of the microbial ecology of complex communities at nearly any scale of interest.
... Comprehensive measurements of molecular com position of samples and identification of unknown bio chemical compounds may eventually become feasible. For example, there have been a few case studies in which machine learning platforms have been trained using Raman spectra of diverse types of pathogenic bacteria and were then able to successfully assign newly meas ured data to closely related groups 106,211,212 . By contrast, Raman microspectroscopy has rarely been applied in microbiology for naive identification or measurement of unknown biochemical compounds. ...
Article
Raman microspectroscopy offers microbiologists a rapid and non-destructive technique to assess the chemical composition of individual live microorganisms in near real time. In this Primer, we outline the methodology and potential for its application to microbiology. We describe the technical aspects of Raman analyses and practical approaches to apply this method to microbiological questions. We discuss recent and potential future applications to determine the composition and distribution of microbial metabolites down to subcellular scale; to investigate the host–microorganism, cell–cell and cell–environment molecular exchanges that underlie the structure of microbial ecosystems from the ocean to the human gut microbiome; and to interrogate the microbial diversity of functional roles in environmental and industrial processes — key themes in modern microbiology. We describe the current technical limitations of Raman microspectroscopy for investigation of microorganisms and approaches to minimize or address them. Recent technological innovations in Raman microspectroscopy will further reinforce the power and capacity of this method for broader adoptions in microbiology, allowing microbiologists to deepen their understanding of the microbial ecology of complex communities at nearly any scale of interest.
... Other hand, spectroscopic methods have also been used in metabolomic platforms; for instance, Fourier transform infra-red (FT-IR) spectroscopy can be applied as an automated and very rapid holistic approach (Winson et al., 1997) providing biomolecular "fingerprints" made up of the vibrational features of microbial cell components (Naumann et al., 1991) and the chemically-based discrimination of intact microbial cells, which may allow for the detection and identification of the most significant groups of biomolecules in a biological system. Such findings may direct the study towards the appropriate omics platforms and analytical methods (Goodacre et al., 1998). In the absence of a universally applicable analytical platform, complementary data can be compiled using several platforms, to allow for better understanding of a biological process and a more informative conclusion to be made (Begley et al., 2009). ...
Article
Full-text available
Introduction Glycerol is a byproduct from the biodiesel industry that can be biotransformed by Escherichia coli to high added-value products such as succinate under aerobic conditions. The main genetic engineering strategies to achieve this aim involve the mutation of succinate dehydrogenase (sdhA) gene and also those responsible for acetate synthesis including acetate kinase, phosphate acetyl transferase and pyruvate oxidase encoded by ackA, pta and pox genes respectively in the ΔsdhAΔack-ptaΔpox (M4) mutant. Other genetic manipulations to rewire the metabolism toward succinate consist on the activation of the glyoxylate shunt or blockage the pentose phosphate pathway (PPP) by deletion of isocitrate lyase repressor (iclR) or gluconate dehydrogenase (gnd) genes on M4-ΔiclR and M4-Δgnd mutants respectively. Objective To deeply understand the effect of the blocking of the pentose phosphate pathway (PPP) or the activation of the glyoxylate shunt, metabolite profiles were analyzed on M4-Δgnd, M4-ΔiclR and M4 mutants. Methods Metabolomics was performed by FT-IR and GC–MS for metabolite fingerprinting and HPLC for quantification of succinate and glycerol. Results Most of the 65 identified metabolites showed lower relative levels in the M4-ΔiclR and M4-Δgnd mutants than those of the M4. However, fructose 1,6-biphosphate, trehalose, isovaleric acid and mannitol relative concentrations were increased in M4-ΔiclR and M4-Δgnd mutants. To further improve succinate production, the synthesis of mannitol was suppressed by deletion of mannitol dehydrogenase (mtlD) on M4-ΔgndΔmtlD mutant that increase ~ 20% respect to M4-Δgnd. Conclusion Metabolomics can serve as a holistic tool to identify bottlenecks in metabolic pathways by a non-rational design. Genetic manipulation to release these restrictions could increase the production of succinate.
... Reading was taken three times for each sample. Data of FTIR spectra were processed according to Goodacre, Timmins [15]. All data were normalized by using Origin Pro Software 2016. ...
... Comprehensive measurements of molecular com position of samples and identification of unknown bio chemical compounds may eventually become feasible. For example, there have been a few case studies in which machine learning platforms have been trained using Raman spectra of diverse types of pathogenic bacteria and were then able to successfully assign newly meas ured data to closely related groups 106,211,212 . By contrast, Raman microspectroscopy has rarely been applied in microbiology for naive identification or measurement of unknown biochemical compounds. ...
Article
Raman microspectroscopy offers microbiologists a rapid and non-destructive technique to assess the chemical composition of individual live microorganisms in near real time. In this Primer, we outline the methodology and potential for its application to microbiology. We describe the technical aspects of Raman analyses and practical approaches to apply this method to microbiological questions. We discuss recent and potential future applications to determine the composition and distribution of microbial metabolites down to subcellular scale; to investigate the host–microorganism, cell–cell and cell–environment molecular exchanges that underlie the structure of microbial ecosystems from the ocean to the human gut microbiomes; and to interrogate the microbial diversity of functional roles in environmental and industrial processes — key themes in modern microbiology. We describe the current technical limitations of Raman microspectroscopy for investigation of microorganisms and approaches to minimize or address them. Recent technological innovations in Raman microspectroscopy will further reinforce the power and capacity of this method for broader adoptions in microbiology, allowing microbiologists to deepen their understanding of the microbial ecology of complex communities at nearly any scale of interest.
... Applications of FT-IR spectroscopic technique are broad in scope, and this was accommodated in the 1990s within the biosciences, as the technique was established to be mostly useful for the discrimination of axenically cultured bacteria (Goodacre et al., 1998), and work on food-related bacteria continues (Lu et al., 2011). The work undertaken during the '90s spawned a considerable amount of interest which led to other areas of research, such as metabolomics, more speci cally referred to as metabolic ngerprinting. ...
Chapter
This chapter reviews different fingerprinting techniques used in food authentication. It shows that the fingerprinting techniques have a wide adoption in food detection, covering most of the food testing aspects. The chapter suggests that fingerprinting technologies have their advantages in food detection: accurate, good reproducibility, low detection limit, and rapid analysis. Fingerprinting technique has become an important technology in food detection and as it is developing rapidly, the applications of fingerprinting technique would be more extended. Spectroscopic and chromatographic fingerprinting techniques in food authentication are absolutely suitable and valid for working in-house when authentication of food samples is required. The food products are usually subjected to a number of commercial treatments, either necessary or desirable, for example, pasteurization or value adding processes or quality assurance processes. Authenticity testing is a quality criterion for food and food ingredients, which is increasingly becoming a result of legislative protection for regional food.
Chapter
Full-text available
Deep neural networks have emerged as a set of robust machine learning tools for computer vision. The suitability of convolutional and recurrent neural networks, along with their variants, is well documented for color image analysis. However, remote sensing and biomedical imaging often rely on hyperspectral images containing more than three channels for pixel-level characterization. Deep learning can facilitate image analysis in multi-channel images; however, network architecture and design choices must be tailored to the unique characteristics of this data. In this two-part series, we review convolution and recurrent neural networks as applied to hyperspectral imagery. Part I focuses on the algorithms and techniques, while Part II focuses on application-specific design choices and real-world remote sensing and biomedical test cases. These chapters also survey recent advances and future directions for deep learning with hyperspectral images.
Chapter
Full-text available
Honey is essentially a highly concentrated aqueous solution of two sugars, dextrose and fructose, and there may be other sugars present as well. Commercially, there are mono-floral honey and poly-floral honey. The identification of honey is very typical because many substances are used in honey as adulterants, like syrups, high fructose corn syrups, invert syrups, or high fructose inulin syrups. Traceability of honey means that National science foundation international provides certification and testing services for honey producers along with food and personal care processors to independently verify the origin traceability of honey. The different fingerprinting techniques have been exploited for authentication and traceability of honey. Nuclear magnetic resonance spectroscopy can also be considered as a fingerprinting technique. Gas chromatography coupled with mass spectrometry is an analytical method used to identify different substances within a test sample. Fluorescence spectroscopy measures the intensity of photons emitted from a sample after it has absorbed photons.
Article
Bakterien der Gattung Bacillus sind im Boden weit verbreitet und können unter ungünstigen Bedingungen Dauerstadien (Sporen) bilden. Bakteriensporen sind sehr widerstandfähig und können lange Zeit in der Umwelt überdauern. Die meisten Bazillen sind harmlos für Mensch und Tier. Zu ihnen gehören aber auch einige bekannte Krankheitserreger wie z.B. Bacillus anthracis (Anthrax, Milzbrand). Diese vergessene Tierkrankheit war einst relativ weit verbreitet. Auch Menschen können sich anstecken. Verschiedene Standorte könnten noch immer mit keimfähigen Sporen kontaminiert sein und so als natürliche Reservoire dienen. Im Rahmen dieser Doktorarbeit wurden historische Recherchen über das Auftreten von Tier-Anthrax im Kanton Zürich durchgeführt und ein Kataster von möglicherweise kontaminierten Standorten erstellt. Weiter wurden verschiedene Methoden zur Detektion und Identifikation von Bekteriensporen entwickelt und getestet, die schnell, einfach in der Anwendung und kostengünstig sein sollten. Dazu gehörten physikalisch- chemische Methoden wie Terbiumfluoreszenz und Fourier Transformation Infrarot (FTIR) Spektroskopie sowie selektive und chromogene Nährmedien. Schliesslich wurde mit Hilfe von molekularbiologischen Methoden (PCR-TTGE fingerprinting Technik) die Diversität von Bacillus Arten im Boden untersucht. B. anthracis konnte in keiner der Proben eines potentiell noch kontaminierten Standortes nachgewiesen werden. Bacteria of the genus Bacillus are common in soil and are able to build dormant stages(endospores) under unfavorable environmental conditions. These spores are extremely resistant to environmental stress and found to survive in soil for many decades. Most Bacilli are harmless to humans and animals, but some of them (e.g. Bacillus anthracis) can cause severe diseases like anthrax. This long-forgotten animal disease once was relatively widespread. It is zoonotic, that means, that it can be transmitted to people. Specific sites could still be contaminated with viable spores and serve as natural reservoirs. Within the scope of this doctoral thesis historical investigations about animal anthrax in the canton of Zurich were carried out and a land register of possibly contaminated locations was established. Furthermore different methods to detect and identify bacterial spores were evaluated and (further) developed. They had to be fast, easy and inexpensive. These were physico-chemical methods such as Terbium fluorescence and Fourier transform infrared (FTIR) spectroscopy as well as selective and chromogenic media. Finally the bacillus diversity in soil was assessed by molecular methods (PCR-TTGE fingerprinting technique). B. anthracis was detected in none of the samples collected from a potentially still contaminated site.
Preprint
We present a computational live bacteria detection system that periodically captures coherent microscopy images of bacterial growth inside a 60 mm diameter agar-plate and analyzes these time-lapsed holograms using deep neural networks for rapid detection of bacterial growth and classification of the corresponding species. The performance of our system was demonstrated by rapid detection of Escherichia coli and total coliform bacteria (i.e., Klebsiella aerogenes and Klebsiella pneumoniae subsp. pneumoniae) in water samples. These results were confirmed against gold-standard culture-based results, shortening the detection time of bacterial growth by >12 h as compared to the Environmental Protection Agency (EPA)-approved analytical methods. Our experiments further confirmed that this method successfully detects 90% of bacterial colonies within 7-10 h (and >95% within 12 h) with a precision of 99.2-100%, and correctly identifies their species in 7.6-12 h with 80% accuracy. Using pre-incubation of samples in growth media, our system achieved a limit of detection (LOD) of ~1 colony forming unit (CFU)/L within 9 h of total test time. This computational bacteria detection and classification platform is highly cost-effective (~$0.6 per test) and high-throughput with a scanning speed of 24 cm2/min over the entire plate surface, making it highly suitable for integration with the existing analytical methods currently used for bacteria detection on agar plates. Powered by deep learning, this automated and cost-effective live bacteria detection platform can be transformative for a wide range of applications in microbiology by significantly reducing the detection time, also automating the identification of colonies, without labeling or the need for an expert.
Chapter
Full-text available
This chapter discusses different analytical approaches for the detection of fruits and vegetables based on their geographical origins. These analytical techniques have been subdivided into four different types based on their principles: mass spectrometry techniques, spectroscopic techniques, separation techniques, and other techniques. Fingerprinting is the combined use of a variety of analytical techniques or methods. Applications of nuclear magnetic resonance spectrometry in food fingerprinting are: analysis of food components, determination of water content, amino acids, and carbohydrates. Fourier-transform infrared spectroscopy is a widely used technique in the determination of adulterants in food material. UV fingerprinting techniques work on the principle of UV absorption spectra and they reflect the transitions of valence electron energy levels of elements. Isotope ratio mass spectrometry works on the principle of identification of chemically identical compounds on the basis of their isotopic contents. Gas chromatograph techniques have been widely applied for the authentication and identification of a wide range of food for quality prediction.
Chapter
Full-text available
This chapter explains the customer satisfaction, improvement in food crises management, improvement in food supply chain management, competence development, technological and scientific contribution, and contribution to agricultural sustainability. Denaturing gradient gel electrophoresis is the most used fingerprinting techniques in food microbiology. The development and implementation of IT-supported food traceability systems (FTS) can improve the operational planning and increase the efficiencies of food logistics processes. Often, a combination of two or more factories influences the development and implementation of FTSs. Many driving forces behind the development and implementation of FTSs have been identified. These drivers have been categorized into five groups: food safety and quality, regulatory, social, economic, and technological concerns. Food quality and safety crises, in turn, cause significant crises for the economic and marketing relationships at the national and international levels. Traceability systems increase the quality of food and food production systems as they increase the awareness of workers through their focus on data capturing and documentation processes.
Article
Full-text available
The Gram-negative bacterial pathogen Campylobacter jejuni is a major cause of foodborne gastroenteritis worldwide. Rapid detection and identification of C. jejuni informs timely prescription of appropriate therapeutics and epidemiological investigations. Here, for the first time, we report the applicability of Raman spectroscopy, surface-enhanced Raman scattering (SERS) and matrix-assisted laser desorption/ionisation mass spectrometry (MALDI-TOF-MS) combined with chemometrics for rapid differentiation and characterisation of mutants of a single isogenic C. jejuni strain that disrupt the production of prominent surface features (capsule, flagella and glycoproteins) of the bacterium. Multivariate analysis of the spectral data obtained from these different physicochemical tools revealed distinctive biochemical differences which consistently discriminated between these mutants. In order to generate biochemical and phenotypic information from different locations in the cell – cell wall versus cytoplasm – we developed two different in situ methods for silver nanoparticle (AgNPs) production, and compared this with simple mixing of bacteria with pre-synthesised AgNPs. This SERS trilogy (simple mixing with premade AgNPs and two in situ AgNPs production methods) presents an integrated platform with potential for rapid, accurate and confirmatory detection of pathogenic bacteria based on cell envelope or intracellular molecular dynamics. Our spectral findings demonstrate that Raman, SERS and MALDI-TOF-MS are powerful metabolic fingerprinting techniques capable of discriminating clinically relevant cell wall mutants of a single isogenic bacterial strain.
Article
Full-text available
Early identification of pathogenic bacteria in food, water, and bodily fluids is very important and yet challenging, owing to sample complexities and large sample volumes that need to be rapidly screened. Existing screening methods based on plate counting or molecular analysis present various tradeoffs with regard to the detection time, accuracy/sensitivity, cost, and sample preparation complexity. Here, we present a computational live bacteria detection system that periodically captures coherent microscopy images of bacterial growth inside a 60-mm-diameter agar plate and analyses these time-lapsed holograms using deep neural networks for the rapid detection of bacterial growth and the classification of the corresponding species. The performance of our system was demonstrated by the rapid detection of Escherichia coli and total coliform bacteria (i.e., Klebsiella aerogenes and Klebsiella pneumoniae subsp. pneumoniae) in water samples, shortening the detection time by >12 h compared to the Environmental Protection Agency (EPA)-approved methods. Using the preincubation of samples in growth media, our system achieved a limit of detection (LOD) of ~1 colony forming unit (CFU)/L in ≤9 h of total test time. This platform is highly cost-effective (~$0.6/test) and has high-throughput with a scanning speed of 24 cm²/min over the entire plate surface, making it highly suitable for integration with the existing methods currently used for bacteria detection on agar plates. Powered by deep learning, this automated and cost-effective live bacteria detection platform can be transformative for a wide range of applications in microbiology by significantly reducing the detection time and automating the identification of colonies without labelling or the need for an expert.
Article
Optical spectroscopy and imaging techniques play important roles in many fields such as disease diagnosis, biological study, information technology, optical science, and materials science. Over the past decade, machine learning (ML) has proved promising in decoding complex data, enabling rapid and accurate analysis of optical spectra and images. This review aims to shed light on various ML algorithms for optical data analysis with a focus on their applications in a wide range of fields. The goal of this work is to sketch the validity of ML‐based optical data decoding. The review concludes with an outlook on unaddressed problems and opportunities in this emerging subject that interfaces optics, data science, and ML.
Article
While Raman spectroscopy can provide label‐free discrimination between highly similar biological species, the discrimination is often marginal, and optimal use of spectral information is imperative. Here we compare two machine learning models, an Artificial Neural Network and a Support Vector Machine for discriminating between Raman spectra of eleven bacterial mutants of Escherichia coli MDS42. While we find that both models discriminate the eleven bacterial strains with similarly high accuracy, sensitivity, and specificity, it is clear that the models form different class boundaries. By extracting strain‐specific (and function‐specific) spectral features utilized by the models, we find that both models utilize a small subset of high intensity peaks while separate subsets of lower intensity peaks are utilized by only one method or the other. This analysis highlights the need for methods to use the complete spectral information more effectively, beginning with a better understanding of the distinct information gained from each model. This article is protected by copyright. All rights reserved.
Article
Pathogens and antimicrobial resistance (AMR) are emerging as major global threats to public health. Wastewater, as the unique interface between environments and humans receiving and spreading pathogens and AMR, is playing a more important role than ever before for monitoring of public health. Here, by pinpointing pathogens and AMR, we reviewed the most recent technological advancements in Raman biosensors (single-cell Raman, Raman-stable isotope probing, surface-enhanced Raman, statistical analysis) and molecular methods (polymerase chain reaction, metagenomics and single-cell genomics) for phenotypic and genotypic surveillance, respectively. In particular, the importance of integrating phenotypic and genotypic analysis via targeted single-cell sorting for a complementary and holistic surveillance and understanding of health risk was highlighted. We further suggest technological requirements to enhance wastewater surveillance and better inform tackling strategy against pathogens and AMR.
Article
Full-text available
Raman spectroscopy has the ability to retrieve molecular information from live biological samples non-invasively through optical means. Coupled with machine learning, it is possible to use this large amount of information to create models that can predict the state of new samples. We study here linear models, whose separation coefficients can be used to interpret which bands are contributing to the discrimination, and compare the performance of principal component analysis coupled with linear discriminant analysis (PCA/LDA), with regularized logistic regression (Lasso). By applying these methods to single-cell measurements for the detection of macrophage activation, we found that PCA/LDA yields poorer performance in classification compared to Lasso, and underestimates the required sample size to reach stable models. Direct use of Lasso (without PCA) also yields more stable models, and provides sparse separation vectors that directly contain the Raman bands most relevant to classification. To further evaluate these sparse vectors, we apply Lasso to a well-defined case where protein synthesis is inhibited, and show that the separating features are consistent with RNA accumulation and protein levels depletion. Surprisingly, when features are selected purely in terms of their classification power (Lasso), they consist mostly of side bands, while typical strong Raman peaks are not present in the discrimination vector. We propose that this occurs because large Raman bands are representative of a wide variety of intracellular molecules and are therefore less suited for accurate classification.
Article
This study developed an in-field analytical technique for food samples by integrating filtration into a surface-enhanced Raman spectroscopy (SERS) microchip. This microchip embedded a filter membrane in the chip inlet to eliminate interfering particulates and enrich target analytes. The design and geometry of the channel were optimised by finite-elemental method (FEM) to tailor variations of flow velocity (within 0–24 μL/s) and facilitate efficient mixing of the filtrate with nanoparticles in two steps. Four pesticides (thiabendazole, thiram, endosulfan, and malathion) were successfully detected either individually or as a mixture in strawberries using this sensor. Strong Raman signals were obtained for the four studied pesticides and their major peaks were clearly observable even at a low concentration of 5 µg/kg. Limits of detection of four pesticides in strawberry extract were in the range of 44–88 μg/kg, showing good sensitivity of the sensor to the target analytes. High selectivity of the sensor was also proved by successful detection of each individual pesticide as a mixture in strawberry matrices. High recoveries (90–122%) were achieved for the four pesticides in the strawberry extract. This sensor is the first filter-based SERS microchip for identification and quantification of multiple target analytes in complex food samples.
Article
Raman spectroscopy is a promising tool for identifying microbial phenotypes based on single cell Raman spectra reflecting cellular biochemical biomolecules. Recent studies using Raman spectroscopy have mainly analyzed phenotypic changes caused by microbial interactions or stress responses (e.g., antibiotics) and evaluated the microbial activity or substrate specificity under a given experimental condition using stable isotopes. Lack of labelling and the nondestructive pretreatment and measurement process of Raman spectroscopy have also aided in the sorting of microbial cells with interesting phenotypes for subsequently conducting physiology experiments through cultivation or genome analysis. In this review, we provide an overview of the principles, advantages, and status of utilization of Raman spectroscopy for studies linking microbial phenotypes and functions. We expect Raman spectroscopy to become a next-generation phenotyping tool that will greatly contribute in enhancing our understanding of microbial functions in natural and engineered systems.
Article
Rapid and accurate classification and discrimination of bacteria is an important task and has been highlighted recently for rapid diagnostics using real-time results. Coupled with a recent report by Jim O'Neill [https://amr-review.org] that if left unaddressed antimicrobial resistance (AMR) in bacteria could kill 10 million people per year by 2050, which would surpass current cancer mortality, this further highlights the need for unequivocal identification of microorganisms. Whilst traditional microbiological testing has offered insights into the characterisation and identification of a wide range of bacteria, these approaches have proven to be laborious and time-consuming and are not really fit for purpose, considering the modern day speed and volume of international travel and the opportunities it creates for the spread of pathogens globally. To overcome these disadvantages, modern analytical methods, such as mass spectrometry (MS) and vibrational spectroscopy, that analyse the whole organism, have emerged as essential alternative approaches. Currently within clinical microbiology laboratories, matrix assisted laser desorption ionisation (MALDI)-MS is the method of choice for bacterial identification. This is largely down to its robust analysis as it largely measures the ribosomes which are always present irrespective of how the bacteria are cultured. However, MALDI-MS requires large amounts of biomass and infrared spectroscopy and Raman spectroscopy are attractive alternatives as these physicochemical bioanalytical techniques have the advantages of being rapid, reliable and cost-effective for analysing various types of bacterial samples, even at the single cell level. In this review, we discuss the fundamental applications, advantages and disadvantages of modern analytical techniques used for bacterial characterisation, classification and identification.
Article
Full-text available
Hyperspectral classification using artificial neural networks is commonly applied on camera dependent interpolated data, or on the results of a dimensionality reduction algorithm. While these methods usually produce satisfactory results, they have severe limitations when part of the spectrum is missing, for example when parts of the image are overexposed or affected by bad pixels. This article presents an input layer based on the Haar transform for artificial neural networks used for hyperspectral data classification. This input layer is designed to perform efficiently with incomplete data and is independent of the specific bands used by the camera. This could enable providing pre-trained neural networks, which can be used with a camera with different specifications than the one used for training. This paper shows that a classifier for mineral identification built using this approach performs better than standard normalization on incomplete spectra, and similarly on complete spectra. Additionally, it shows that such a classifier matches local spectral features, and therefore that the artificial neural network is matching the spectrum shape.
Article
Full-text available
Bacterial infection is a global burden that results in numerous hospital visits and deaths annually. The rise of multi-drug resistant bacteria has dramatically increased this burden. Therefore, there is a clinical need to detect and identify bacteria rapidly and accurately in their native state or a culture-free environment. Current diagnostic techniques lack speed and effectiveness in detecting bacteria that are culture-negative, as well as options for in vivo detection. The optical detection of bacteria offers the potential to overcome these obstacles by providing various platforms that can detect bacteria rapidly, with minimum sample preparation, and, in some cases, culture-free directly from patient fluids or even in vivo. These modalities include infrared, Raman, and fluorescence spectroscopy, along with optical coherence tomography, interference, polarization, and laser speckle. However, these techniques are not without their own set of limitations. This review summarizes the strengths and weaknesses of utilizing each of these optical tools for rapid bacteria detection and identification.
Article
Full-text available
Through this pilot study, the association between Raman spectroscopy and Machine Learning algorithms were used for the first time with the purpose of distillates differentiation with respect to trademark, geographical and botanical origin. Two spectral Raman ranges (region I—200–600 cm⁻¹ and region II—1200–1400 cm⁻¹) appeared to have the higher discrimination potential for the investigated distillates. The proposed approach proved to be a very effective one for trademark fingerprint differentiation, a model accuracy of 95.5% being obtained (only one sample was misclassified). A comparable model accuracy (90.9%) was achieved for the geographical discrimination of the fruit spirits which can be considered as a very good one taking into account that this classification was made inside Transylvania region, among neighbouring areas. Because the trademark fingerprint is the prevailing one, the successfully distillate type differentiation, with respect to the fruit variety, was possible to be made only inside of each producing entity.
Article
Full-text available
Dog erythrocytic membrane antigen plays a major role for determining blood group. Structural and molecular characterization of erythrocytic membrane antigen improves the production of blood typing antisera, to study the auto antibody production in canine autoimmune haemolytic anemia. The proteins in the lipid domain arranged from the inside of the erythrocyte to the outside. The integral membrane proteins include membrane protein 3 visible in Coomassie Brilliant Blue-stained polyacrylamide gels. The erythrocyte cytoskeleton consists of spectrin, ankyrin, actin and protein 4.1 form a filamentous network under the lipid bilayer of erythroctic membrane. Most characteristic bands are associated with the CO =NH group referred to as amide A have NH stretching mostly found at 3500 cm-1 wave number. The amide B had NH stretching found at the region of 3100 cm-1 and amide I & III were used to estimate the secondary structure of proteins. Amide I mode which ranges from 1580 cm-1 to 1700 cm-1 and very sensitive to the backbone conformation and not affected by the side chains. Amide I band can be de convoluted with various sub-bands which directly correlate with various secondary structures The good intensity sharp peak at the level of 1468 cm-1 and 2034.694 cm-1 were taken for 2D analysis the CCD cts scale bar shows differences between DEA1.1 positive and negative dog erythrocytic membrane antigen.
Chapter
Full-text available
Wine control is traditionally and strongly associated with the proof of authenticity. Wine authenticity is very important, especially in case of quality control and consumer information. Since wine quality is dependent on the consumer demands, compliance with traceability provisions satisfies the associated economic needs. Traceability in the wine industry has an important role in the quality assurance management system. Grape growers are responsible for the production, harvest, and delivery of grapes. Wine producers are responsible for the production, manufacture, and/or blending of wine products. There are a variety of techniques and processes of wine-making, authorized or used fraudulently, altering appreciably the compositional and sensorial wine parameters. This aspect of authenticity has a particular relevance to special wines like the sparkling wine, the oxidative wines. Investigation of mineral elements in wine is the main procedure to authenticate the geographical origins of wines.
Article
Full-text available
The application of wavelet denoising to infrared spectra was investigated, Six different wavelet denoising methods (SURE, VISU, HYBRID, MINMAX, MAD and WAVELET PACKETS) were applied to pure infrared spectra with various added levels of homo- and heteroscedastic noise, The performances of the wavelet denoising methods were compared with the standard Fourier and moving mean filtering in terms of root mean square errors between the pure and denoised spectra and visual quality of the denoised spectrum, The use of predictive ability as a possible objective criterion for denoising performance was also investigated, The main conclusion is that for very low signal-to-noise ratios (SIN) the standard denoising methods (Fourier and moving mean) are comparable to the more sophisticated methods, At higher SIN levels the wavelet denoising methods, in particular the HYBRID and VISU methods, are better, Wavelet methods are also better in restoring the visual quality of the denoised infrared spectra.
Article
The most fundamental questions such as whether a cell is alive, in the sense of being able to divide or to form a colony, may sometimes be very hard to answer, since even axenic microbial cultures are extremely heterogeneous. Analyses that seek to correlate such things as viability, which is a property of an individual cell, with macroscopic measurements of culture variables such as ATP content, respiratory activity, and so on, must inevitably fail. It is therefore necessary to make physiological measurements on individual cells. Flow cytometry is such a technique, which allows one to analyze cells rapidly and individually and permits the quantitative analysis of microbial heterogeneity. It therefore offers many advantages over conventional measurements for both routine and more exploratory analyses of microbial properties. While the technique has been widely applied to the study of mammalian cells, is use in microbiology has until recently been much more limited, largely because of the smaller size of microbes and the consequently smaller optical signals obtainable from them. Since these technical barriers no longer hold, flow cytometry with appropriate stains has been used for the rapid discrimination and identification of microbial cells, for the rapid assessment of viability and of the heterogeneous distributions of a wealth of other more detailed physiological properties, for the analysis of antimicrobial drug-cell interactions, and for the isolation of high-yielding strains of biotechnological interest. Flow cytometric analyses provide an abundance of multivariate data, and special methods have been devised to exploit these. Ongoing advances mean that modern flow cytometers may now be used by nonspecialists to effect a renaissance in our understanding of microbial heterogeneity.
Article
Surface-enhanced Raman scattering (SERS) spectroscopy is now a well-established phenomenon, which has been thoroughly characterized in a variety of interfacial and colloidal environments. Although some quantitative aspects of the underlying enhancement mechanisms apparently remain unresolved, attention is now shifting towards application of SERS to explore phenomena of chemical, physical, biological and industrial significance. The goal of this review is to appreciate the industrial value of innovative SERS technique on the basis of our experience in development of new SERS-active substrates and in their biomedical and biotechnological applications. Examples of diverse SERS analytical applications as well as some very recent facilities, such as SERS microprobe analysis, SERS fiber optics probes, FT-SERS spectroscopy, SERS detection for high-performance liquid chromatography, etc., are also discussed.
Article
In attempting to analyze, on digital computers, data from basically continuous physical experiments, numerical methods of performing familiar operations must be developed. The operations of differentiation and filtering are especially important both as an end in themselves, and as a prelude to further treatment of the data. Numerical counterparts of analog devices that perform these operations, such as RC filters, are often considered. However, the method of least squares may be used without additional computational complexity and with considerable improvement in the information obtained. The least squares calculations may be carried out in the computer by convolution of the data points with properly chosen sets of integers. These sets of integers and their normalizing factors are described and their use is illustrated in spectroscopic applications. The computer programs required are relatively simple. Two examples are presented as subroutines in the FORTRAN language.
Article
Unselected urinary pathogens from general practice and hospital have been tested for sensitivity to a range of antimicrobial agents for the last 22 years. There have been substantial changes. In general practice there has been a considerable increase in the proportion of staphylococcal infections from 5.1% to a peak of 14.8% in 1982 and a more recent decline to 4.0%. There has also been a decrease in the proportion caused by Proteus mirabilis, from 9.2% to 4.3%. Similar, but smaller, changes have been observed in the proportions of hospital urinary tract infections caused by these organisms, while the proportion of hospital infections due to Klebsiella spp. and Enterobacter spp. has fallen from 16.8% to 7.3%
Article
The next generation of satellite-borne sensors will combine high spatial resolution with fine spectral resolution. A typical data set for a single frame of imagery may contain a few hundred images occupying many gigabytes of space. Clearly, traditional image processing algorithms cannot be directly applied to such a vast quantity of data. We investigate enhancement and compression algorithms that use the spectral correlation present in high-resolution imagery to reduce the computational complexity of processing the imagery. The algorithm employs a principal component transformation to reduce the size of the data set. Enhancing the reduced set of images provides equivalent results to processing each of the original images with far fewer computations. The compression algorithm utilizes a hybrid discrete cosine transform- differential pulse code modulation (DCT-DPCM) transform. The DCT is computed for each image, a bit map is generated for the DCT coefficients, and DPCM is used to encode the coefficients across the bands. Compression at less than 0.5 bits/pixel with negligible visual degradation is obtained.
Article
A perceptual-based multiresolution image fusion technique is demonstrated using the Airborne Visible and Infrared Imaging Spectrometer (AVIRIS) hyperspectral sensor data. The AVIRIS sensor, which simultaneously collects information in 224 spectral bands that range from 0.4 to 2.5 micrometers in approximately 10-nm increments, produces 224 images, each representing a single spectral band. The fusion algorithm consists of three stages. First, a Daubechies orthogonal wavelet basis set is used to perform a multiresolution decomposition of each spectral image. Next, the coefficients from each image are combined using a perceptual-based weighting. The weighting of each coefficient, from a given spectral band image, is determined by the spatial-frequency response (contrast sensitivity) of the human visual system. The spectral image with the higher saliency value, where saliency is based on a perceptual energy, will receive the larger weight. Finally, the fused coefficients are used for reconstruction to obtain the fused image. The image fusion algorithm is analyzed using test images with known image characteristics and image data from the AVIRIS hyperspectral sensor. By analyzing the signal-to-noise ratios and visual aesthetics of the fused images, contrast-sensitivity-based fusion is shown to provide excellent fusion results and to outperform previous fusion methods.
Article
We demonstrate the use of a single-stage spectrograph and charge-coupled-device (CCD) detector to collect near-infrared (NIR) Raman spectra from intact human arterial tissue. With 810-nm excitation, the fluorescence emission from human artery is sufficiently weak to allow observation of Raman bands more rapidly with the spectrograph/CCD system than with 1064-nm excited FT-Raman systems. We also present a method for removing the broad-band emission from the spectra by computing the difference of two emission spectra collected at slightly different excitation frequencies. Our results indicate that NIR Raman spectra can be collected in under one second with the spectrograph/CCD system and an optical fiber probe, opening the possibility of in vivo clinical applications.
Article
Confocal Raman microspectroscopy has previously used pinholes placed at the back focal plane of the microscope to provide depth resolution along the optical axis. The process of optimizing the pinhole alignment can often be difficult and time-consuming. We demonstrate a different approach to setting up a confocal Raman microscope using a stigmatic spectrograph and a CCD detector. This arrangement is easy to use and provides a depth resolution of ∼2 μm.
Article
Publisher Summary This chapter reviews the group frequencies in terms of the spectral regions in which they occur. The first part of the chapter outlines an orderly procedure for the initial interpretation of an unknown infrared radiation (IR) spectrum by regions. Then, spectra-structure correlations are shown in a chart form where one can look for the groups that absorb in a given region or the regions where a given group absorbs. One of the useful features of infrared spectroscopy is its ability to give information about mixtures. When more than one component is present, the spectra tend to be additive but not completely so because of possible mutual interaction, such as hydrogen bonding for example. If the main component in the spectrum has been identified, a comparison with a reference spectrum may reveal some extra bands in the sample spectrum not in the reference. The chapter also presents some selected Infrared spectra and Raman spectra and illustrates their functional group frequencies.
Article
Rapid and reliable methods for the detection and identification of microorganisms are very important. Fortunately, many effective means for bacterial identification have been developed. On the other hand, only a few of these are rapid, accurate, and cost effective. In most instances, if general examinations of bacteria must be performed, presently used techniques are very slow and may not be helpful unless extremely tedious procedures are followed. Traditional methods, which are based upon visual microscopic examination, biochemical reactions, and physiological functions of bacteria, are inherently slow, time consuming, and tedious. Even today, important decisions related to the presence of pathogens have to be made before the results of microbiological tests are available. This situation is changing rapidly, however.
Article
Many neural network classifiers provide outputs which estimate Bayesian a posteriori probabilities. When the estimation is accurate, network outputs can be treated as probabilities and sum to one. Simple proofs show that Bayesian probabilities are estimated when desired network outputs are 1 of M (one output unity, all others zero) and a squared-error or cross-entropy cost function is used. Results of Monte Carlo simulations performed using multilayer perceptron (MLP) networks trained with backpropagation, radial basis function (RBF) networks, and high-order polynomial networks graphically demonstrate that network outputs provide good estimates of Bayesian probabilities. Estimation accuracy depends on network complexity, the amount of training data, and the degree to which training data reflect true likelihood distributions and a priori class probabilities. Interpretation of network outputs as Bayesian probabilities allows outputs from multiple networks to be combined for higher level decision making, simplifies creation of rejection thresholds, makes it possible to compensate for differences between pattern class probabilities in training and test data, allows outputs to be used to minimize alternative risk functions, and suggests alternative measures of network performance.
Article
An abstract is not available.
Article
We propose a network architecture which uses a single internal layer of locally-tuned processing units to learn both classification tasks and real-valued function approximations (Moody and Darken 1988). We consider training such networks in a completely supervised manner, but abandon this approach in favor of a more computationally efficient hybrid learning method which combines self-organized and supervised learning. Our networks learn faster than backpropagation for two reasons: the local representations ensure that only a few units respond to any given input, thus reducing computational overhead, and the hybrid learning rules are linear rather than nonlinear, thus leading to faster convergence. Unlike many existing methods for data analysis, our network architecture and learning rules are truly adaptive and are thus appropriate for real-time use.
Article
The identification of microorganisms using conventional procedures is a highly developed routine employing a wide range of morphological, serological, nutritional and biochemical tests. The application of Py-MS to microbiology remains underdeveloped and ill-defined. The advent of low-cost instrumentation should increase the access of microbiologists to the method. Perhaps the most significant obstacles to the more widespread use of Py- MS are the lack of understanding about what pyrolysis mass spectra represent and why they can be used to differentiate organisms. Detailed studies of microbial pyrolysis products by Py-GC-MS techniques are required to identify the cell constituents that contribute to the spectra and hence the differentiation of microorganisms. In the future, Py-MS should be used in tandem with other chemotaxonomic techniques, so that an improved understanding of the chemical basis of differences between spectra can be acquired.
Article
Pyrolysis mass spectra of complex biological samples can be interpreted for partial chemical composition, even when exact reference spectra of the components involved are not available. The method is based on a factor analysis method: principal component analysis followed by discriminant analysis and graphical rotation. Applications of the procedure are discussed for various sets of samples, viz., fungi (taxonomy), bacteria (detection of capsular polysaccharlde), and human bile (differentiation of cholelithiasis and cholecystitis patients and normals).
Article
A new thin-layer chromatography (TLC) technique termed microchannel TLC, is described. Channels, having the dimensions 400 mu m X 200 mu m X 5 cm (W X D X L), have been packed with a zirconia stationary phase and used for TLC. Subsequent infrared microspectroscopic detection of organic dyes separated in these channels provided excellent diffuse reflectance spectra and an improvement in the minimum identifiable quantity by a factor of about 500 times over previous TLC work using microscope slides. This technique also requires smaller amounts of sample and stationary phase compared to conventional TLC techniques. A practical application of the separation and identification of four polyaromatic hydrocarbons (two of which are isomers) is also shown.
Article
This manuscript describes methods for detecting and modeling nonlinear regions of spectral response in multivariate, multicomponent spectroscopic assays. Simulated data and experimental UV/visible data were used to study the capability of multivariate linear models to approximate nonlinear response. The sources of real and apparent nonlinearity simulated included nonlinear instrument response functions (e.g. stray light), concentration-dependent wavelength shifts, and concentration dependent absorption bandwidth changes. A weighting algorithm was devised to reduce the influence of nonlinear spectral regions in principal component regression (PCR) calibrations, thereby improving the performance of multivariate linear calibration models. Second-order calibration methods using quadratic principal component scores and nonlinear calibration methods using artificial neural networks were compared to unweighted and weighted linear calibration methods. Orthogonal transformation of the input variables was used to significantly improve neural network training speed and reduce calibration error. Some conditions where second-order and nonlinear calibration techniques outperform linear calibration techniques have been identified and are described.
Article
Binary mixtures of the protein lysozyme with glycogen, of DNA or RNA in glycogen, and the tertiary mixture of cells of the bacteria Bacillus subtilis, Escherichia coli, and Staphylococcus aureus were subjected to pyrolysis mass spectrometry. To analyze the pyrolysis mass spectra so as to obtain quantitative information representative of the complex components of the mixtures, partial least-squares regression (PLS), principal components regression (PCR), and fully interconnected feedforward artificial neural networks (ANNs) were studied. in the latter case, the weights were modified using the standard back-propagation algorithm, and the nodes used a sigmoidal squashing function. It was found that each of the methods could be used to provide calibration models which gave excellent predictions for the concentrations of determinands in samples on which they had not been trained. Neural networks were found to provide the most accurate predictions. We also report that scaling the individual nodes on the input layer of ANNs significantly decreased the time taken for the ANNs to learn. Removing masses of low intensity, which perhaps mainly contributed noise to the pyrolysis mass spectra, had little effect on the accuracy of the ANN predictions though could dramatically speed up the learning process (by more than 100-fold) and slightly improved the accuracy of PLS calibrations.
Article
The first part of this review is devoted to medical applications of Raman spectroscopy as a diagnostic or analytical tool. Studies of human arteries, ocular lenses, living cells and chromosomes are reviewed, in addition to recent advantages in cancer diagnostics using Raman spectroscopy. The second and major part is devoted to a relatively new field, surface-enhanced Raman scattering (SERS) spectroscopy of biomedical species. The SERS effect is accompanied by strong quenching of fluorescence and so enabes the range of species that can be investigated by Raman technique to be extended. The ultra-high sensitivity of SERS enables spectra to be obtained at concentrations down to 10−10 M. A wide range of experiments designed to probe the structure, topology and composition of biomedical species using SERS spectroscopy can be envisioned. Some of these currently being studied are: the determination of the distribution of drugs within a living cell and on the cellular membrane, the selective study of cell membrane components, the analysis of crude biomedical mixtures and extracts and new techniques, based on SERS spectroscopy, Fourier transform SERS spectroscopy and the SERS microprobe method.
Article
This review summarizes the literature relating to the application of surface-enhanced Raman scattering (SERS) and surface-enhanced resonance Raman scattering (SERRS) techniques to the study of biological molecules. The emphasis is on publications that have appeared during the period from 1985 to 1991. The review is divided into six major parts. First, a brief overview of the current understanding of the mechanistic aspects of SERS/SERRS is given, with an emphasis on the relationship between theoretical predictions and experimental results. The most common experimental systems (colloids, metal island films, and electrodes) are described. Studies of biological systems are described in the second (small molecules), third (DNA and proteins) and fourth (membranes proteins and membrane preparations) sections. In the fifth or conclusion section, the potential use of SERS or SERRS as a method for obtaining spectra of native biological molecules is evaluated. Finally, the sixth section describes advances in Raman instrumentation in terms of their possible impact on future applications of SERS/SERRS techniques to biological molecules.
Article
Verocytotoxin (VT)-producing and VT-non-producing strains of Escherichia coli of four different serogroups were characterised by pyrolysis-mass spectrometry (Py-MS). Py-MS spectral data were used to train artificial neural networks (ANNs) which then accurately assessed the VT-production status of fresh clinical isolates of E. coli of the same serogroups from their Py-MS spectral data.Serogroup-specific ANNs could be trained successfully with Py-MS data from only one exemplar each of VT-producing and VT-non-producing strains and training was accomplished in less than 1 min. Where more than one VT-producing phenotype occurred within a serogroup it was necessary to include an example of each phenotype in the training set.The combination of Py-MS with ANNs may be an important new, powerful and very rapid method for the detection of a particular biological character, such as exotoxin production within strains of single species or sub-species of bacteria.
Article
Raman spectroscopy is steadily growing in importance in the industrial analytical laboratory. The nature of the equipment, its ease of use and its cost are making the method more acceptable to non-experts. This paper illustrates the capabilities of a recent commercialized Raman system which fulfils the criteria of ease of use, speed and low cost. The system is designed for Raman microscopy and uses a single spectrograph and CCD detector. The combination provides confocal microscopy, high throughput and optimum sensitivity, to the extent that it only requires the use of a low-powered laser to provide high-quality spectral data. The additional feature of direct Raman imaging is seen as being beneficial in the industrial environment in that it provides spatial information over large surface areas quickly and without the need for excessive amounts of data processing. The performance of the instrument is illustrated with applications taken from the industrial environment.
Article
Curie-point pyrolysis mass spectra were obtained from a variety of extra-virgin olive oils, prepared from various cultivars using several mechanical treatments. Some of the oils were adulterated (according to a double-blind protocol) with different amounts of seed oils (50–500 ml of soya, sunflower, peanut, corn or rectified olive oils per litre of mixed oil). Canonical variates analysis indicated that the major source of variation between the pyrolysis mass spectra was due to differences between the cultivars. rather than whether the oils had been adulterated. However, artificial neural networks could be trained (using the back-propagation algorithm) successfully to distinguish virgin oils from those which had been adulterated.
Article
This paper takes abroad, pragmatic view of statistical inference to include all aspects of model formulation. The estimation of model: parameters traditionally assumes that a model has a prespecified known form and takes no account of possible uncertainty regarding the model structure. This implicitly assumes the existence of a 'true' model, which many would regard-as a fiction. In practice model uncertainty is a fact of life and likely to be more serious than other sources of uncertainty which have received far more attention from statisticians. This is true whether the model is specified on subject-matter grounds or, as is increasingly the case, when a model is formulated, fitted and checked on the same data set in an iterative, interactive way. Modern computing power allows a large number of models to be considered and data-dependent specification searches have become the norm in many areas of statistics. The term data mining may be used in this context when the analyst goes to great lengths to obtain a good fit. This paper reviews the effects of model uncertainty, such as too narrow prediction intervals, and the non-trivial biases in parameter estimates which can follow data-based modelling. Ways of assessing and overcoming the effects of model uncertainty are discussed, including the use of simulation and resampling methods, a Bayesian model averaging approach and collecting additional data wherever possible. Perhaps the main aim of the paper is to ensure that statisticians are aware of the problems and start addressing the issues even if there is no simple, general theoretical fix.
Article
This paper is concerned with the representation of a multivariate sample of size n as points P1, P2, …, Pn in a Euclidean space. The interpretation of the distance Δ(Pi, Pj) between the ith and jth members of the sample is discussed for some commonly used types of analysis, including both Q and R techniques. When all the distances between n points are known a method is derived which finds their co-ordinates referred to principal axes. A set of necessary and sufficient conditions for a solution to exist in real Euclidean sapce is found. Q and R techniques are defined as being dual to one another when they both lead to a set of n points with the same inter-point distances. Pairs of dual techniques are derived. In factor analysis the distances between points whose co-ordinrates are the estimated factor scores can be interpreted as D2 with a singular dispersion matrix.
Article
There is much interest in the exploitation of more-or-less sophisticated mathematical methods combining the signals from different sensors for measured variables for estimating the present or future state of bioprocesses. In many cases, however, the application of these methods has failed to take into account many important principles. We therefore summarize what we consider to be some of the key issues, and present them in the form of a guide for assisting the development of these methods and their integration into mainstream biotechnology.
Article
The combination of pyrolysis mass spectrometry (PyMS) and artificial neural networks (ANNs) can be used to quantify levels of penicillins in strains of Penicillium chrysogenum and ampicillin in spiked samples of Escherichia coli. Four P. chrysogenum strains (NRRL 1951, Wis Q176, P1, and P2) were grown in submerged culture to produce penicillins, and fermentation samples were taken aseptically and subjected to PyMS. To deconvolute the pyrolysis mass spectra so as to obtain quantitative information on the titre of penicillins, fully-interconnected feedforward artificial neural networks (ANNs) were studied; the weights were modified using the standard back-propagation algorithm, and the nodes used a sigmoidal squashing function. In addition the multivariate linear regression techniques of partial least squares regression (PLS), principal components regression (PCR) and multiple linear regression (MLR) were applied. The ANNs could be trained to give excellent estimates for the penicillin titre, not only from the spectra that had been used to train the ANN but more importantly from previously unseen pyrolysis mass spectra. All the linear regression methods failed to give accurate predictions, because of the very variable biological backgrounds (the four different strains) in which penicillin was produced and also of the inability of models using linear regression accurately to map non-linearities. Comparisons of squashing functions on the output nodes of identical 150-8-1 neural networks revealed that networks employing linear functions gave more accurate estimates of ampicillin in E. coli near the edges of the concentration range than did those using sigmoidal functions. It was also shown that these neural networks could be successfully used to extrapolate beyond the concentration range on which they had been trained. PyMS with the multivariate clustering technique of principal components analysis was able to differentiate between four strains of P. chrysogenum studied, and was also able to detect phenotypic differences at five, seven, nine or 11 days growth. A crude smapling procedure consisting of homogenised agar plugs proved applicable for rapid analysis of a large number of samples.
Article
The general principle of parsimonious data modeling states that if two models in some way adequately model a given set of data, the one that is described by a fewer number of parameters will have better predictive ability given new data. This concept is of interest in multivariate calibration since several new non-linear modeling techniques have become available. Three such methods are neural networks, projection pursuit regression (PPR) and multivariate adaptive regression splines (MARS). These methods, while capable of modeling non-linearities, typically have very many parameters that need to be estimated during the model building phase. The biased calibration methods, principal components regression (PCR) and partial least squares (PLS) are linear methods and so may not as efficiently describe some types of non-linearities, however have comparably very few parameters to be estimated. It is therefore of interest to study the parsimony principle formally in order to understand under what circumstances the various methods are appropriate. In this paper, the mathematical theory of parsimonious data modeling is presented. The assumptions made in the theory are shown to hold for multivariate calibration methods. This theory is used to provide a procedure for selecting the most parsimonious model structure for a given calibration application.
Article
A new approach founded on Radial Basis Functions (RBF) and Partial Least Squares (PLS) is proposed to model non-linear chemical systems. Its performance is demonstrated for two simulated examples and compared with those of Multilayer Feedforward Network (MLP), Radial Basis Function Network (RBFN), and Spline-PLS. Good performance and a guaranteed learning algorithm of the RBF-PLS approach makes it an attractive alternative for the earlier established methods.
Article
Raman spectroscopy is being used to study biological molecules for some three decades now. Thanks to continuing advances in instrumentation more and more applications have become feasible in which molecules are studied in situ, and this has enabled Raman spectroscopy to enter the realms of biomedicine and cell biology [1–5].Here we will describe some of the recent work carried out in our laboratory, concerning studies of human white blood cells and further instrumentational developments.
Article
We introduce diffuse-reflectance absorbance spectroscopy in the mid-infrared as a novel method of chemical imaging for the rapid screening of biological samples for metabolite overproduction, using mixtures of ampicillin with Escherichia coli and Staphylococcus aureus as model systems. Deconvolution of the hyperspectral information provided by the raw diffuse reflectance-absorbance mid-infrared spectra was achieved using a combination of principal components analysis (PCA), artificial neural networks (ANNs) and partial least squares regression (PLS). Whereas a univariate approach necessitates appropriate data selection to remove any interferences, the chemometrics/hyperspectral approach could be employed to permit filtering of undesired components to give accurate quantification by PLS and ANNs without any preprocessing. The use of PCs as inputs to the ANNs decreased the training time from some 12 h to ca. 5 min. Equivalent concentrations of ampicillin between 0.05 and 20 mM in an E. coli or S. aureus background were quantified with >95% accuracy using this approach.
Article
A 10% (w/v) beet sucrose solution was used to adulterate freshly squeezed orange juice over the range 0–20% (or 0–20 g l−1 of added sucrose). Samples were analysed by the rapid automated screening technique of Curie-point pyrolysis mass spectrometry (PyMS). To deconvolve these spectra neural cognition-based methods of multilayer perceptrons (MLPs) and radial basis functions (RBFs) and the linear regression technique of partial least squares (PLS) were studied. It was found that each of the methods could be used to provide calibration models which gave excellent predictions for the level of sucrose adulteration at levels below 1% for samples, with an accuracy of ± 1.3%, on which they had not been trained. The best results were obtained using PLS when 8 latent variables were employed for predictions. Furthermore, the inputs to MLPs could be reduced using principal components analysis (PCA) from 150 masses to 8 PC scores without any deterioration of the predictive ability of the model, highlighting that PCA is an excellent pre-processing step which has the potential to speed up neural network learning as there are fewer weights to update. Since any foodstuff can be pyrolysed in this way, the combination of PyMS with chemometrics constitutes a rapid, powerful and novel approach to the quantitative assessment of food adulteration generally.
Article
Pyrolysis mass spectrometry is a rapid and high-resolution method for the analysis of otherwise non-volatile material and has been widely applied for discriminating between closely related microbial strains. Recent advances in statistical and neural network methods based on supervised learning have now permitted exploitation of pyrolysis mass spectrometry in the quantitative analysis of many diverse samples of biotechnological interest; the technique may thus be regarded as an ‘anything-sensor’.
Article
There have been several recent studies concerning feedforward networks and the problem of approximating arbitrary functionals of a finite number of real variables. Some of these studies deal with cases in which the hidden-layer nonlinearity is not a sigmoid. This was motivated by successful applications of feedforward networks with nonsigmoidal hidden-layer units. This paper reports on a related study of radial-basis-function (RBF) networks, and it is proved that RBF networks having one hidden layer are capable of universal approximation. Here the emphasis is on the case of typical RBF networks, and the results show that a certain class of RBF networks with the same smoothing factor in each kernel node is broad enough for universal approximation.