Development of robust calibration models using support vector machines for spectroscopic monitoring of blood glucose

Raman-based PAT for VLP precipitation: systematic data diversification and preprocessing pipeline identification

Virus-like particles (VLPs) are a promising class of biopharmaceuticals for vaccines and targeted delivery. Starting from clarified lysate, VLPs are typically captured by selective precipitation. While VLP precipitation is induced by step-wise or continuous precipitant addition, current monitoring approaches do not support the direct product quantification, and analytical methods usually require various, time-consuming processing and sample preparation steps. Here, the application of Raman spectroscopy combined with chemometric methods may allow the simultaneous quantification of the precipitated VLPs and precipitant owing to its demonstrated advantages in analyzing crude, complex mixtures. In this study, we present a Raman spectroscopy-based Process Analytical Technology (PAT) tool developed on batch and fed-batch precipitation experiments of Hepatitis B core Antigen VLPs. We conducted small-scale precipitation experiments providing a diversified data set with varying precipitation dynamics and backgrounds induced by initial dilution or spiking of clarified Escherichia coli-derived lysates. For the Raman spectroscopy data, various preprocessing operations were systematically combined allowing the identification of a preprocessing pipeline, which proved to effectively eliminate initial lysate composition variations as well as most interferences attributed to precipitates and the precipitant present in solution. The calibrated partial least squares models seamlessly predicted the precipitant concentration with R 2 of 0.98 and 0.97 in batch and fed-batch experiments, respectively, and captured the observed precipitation trends with R 2 of 0.74 and 0.64. Although the resolution of fine differences between experiments was limited due to the observed non-linear relationship between spectral data and the VLP concentration, this study provides a foundation for employing Raman spectroscopy as a PAT sensor for monitoring VLP precipitation processes with the potential to extend its applicability to other phase-behavior dependent processes or molecules.

Is Raman the best strategy towards the development of non-invasive continuous glucose monitoring devices for diabetes management?

Diabetes has no well-established cure; thus, its management is critical for avoiding severe health complications involving multiple organs. This requires frequent glycaemia monitoring, and the gold standards for this are fingerstick tests. During the last decades, several blood-withdrawal-free platforms have been being studied to replace this test and to improve significantly the quality of life of people with diabetes (PWD). Devices estimating glycaemia level targeting blood or biofluids such as tears, saliva, breath and sweat, are gaining attention; however, most are not reliable, user-friendly and/or cheap. Given the complexity of the topic and the rise of diabetes, a careful analysis is essential to track scientific and industrial progresses in developing diabetes management systems. Here, we summarize the emerging blood glucose level (BGL) measurement methods and report some examples of devices which have been under development in the last decades, discussing the reasons for them not reaching the market or not being really non-invasive and continuous. After discussing more in depth the history of Raman spectroscopy-based researches and devices for BGL measurements, we will examine if this technique could have the potential for the development of a user-friendly, miniaturized, non-invasive and continuous blood glucose-monitoring device, which can operate reliably, without inter-patient variability, over sustained periods.

Sense and Learn: Recent Advances in Wearable Sensing and Machine Learning for Blood Glucose Monitoring and Trend-Detection

Diabetes mellitus is characterized by elevated blood glucose levels, however patients with diabetes may also develop hypoglycemia due to treatment. There is an increasing demand for non-invasive blood glucose monitoring and trends detection amongst people with diabetes and healthy individuals, especially athletes. Wearable devices and non-invasive sensors for blood glucose monitoring have witnessed considerable advances. This review is an update on recent contributions utilizing novel sensing technologies over the past five years which include electrocardiogram, electromagnetic, bioimpedance, photoplethysmography, and acceleration measures as well as bodily fluid glucose sensors to monitor glucose and trend detection. We also review methods that use machine learning algorithms to predict blood glucose trends, especially for high risk events such as hypoglycemia. Convolutional and recurrent neural networks, support vector machines, and decision trees are examples of such machine learning algorithms. Finally, we address the key limitations and challenges of these studies and provide recommendations for future work.

Tissue Phantoms for Biomedical Applications in Raman Spectroscopy: A Review

Raman spectroscopy is a group of analytical techniques, currently applied in several research fields, including clinical diagnostics. Tissue-mimicking optical phantoms have been established as an essential intermediate stage for medical applications with their employment from spectroscopic techniques to be constantly growing. This review outlines the types of tissue phantoms currently employed in different biomedical applications of Raman spectroscopy, focusing on their composition and optical properties. It is therefore an attempt to present an informed range of options for potential use to the researchers.

Non-invasive assessments of the advanced glycation end products in human skin using reflectance NIR spectroscopy

Advanced glycation end products (AGE) are produced by non-enzymatic reaction between glucose and biomolecules including proteins. AGE accumulation is known to cause alternations of structure and function in proteins and to be related with an increased risk of diabetic complications, cardiovascular diseases, and aging processes. Conventionally, AGE accumulation has been estimated by measuring auto fluorescence level using ultraviolet (UV) light excitation. In this study, we investigated an alternative approach to estimate auto fluorescence level and thus AGE accumulation in in vivo human skin using NIR (Near-Infrared) spectroscopy. To examine spectral features attributed to glycation in proteins, we first analyzed in vitro NIR spectra from native and glycated protein. Then, we further examined NIR spectra of in vivo skin from human subjects, and estimated their auto fluorescence level using several multivariate regression approaches. Our analysis in in vitro spectra from native and glycated albumin revealed that glycation may affect -CH and -NH stretching. Furthermore, we elucidated that those bands for -CH and -NH may be responsible for the variation in auto fluorescence level in human skin NIR spectra. Finally, auto fluorescence level was estimated from those NIR spectra using several multivariate regression methods: principal component regression (PCR), partial least square regression (PLS-R) and support vector regression (SVR). Among the three methods, SVR showed the best performance. We demonstrated in this study that NIR spectroscopy can be used as an alternative non-invasive method to estimate AGE accumulation in in vivo human skin tissue without UV radiation on skin tissue.

A Layer-by-Layer Approach To Retain a Fluorescent Glucose Sensing Assay within the Cavity of a Hydrogel Membrane

A continuous glucose monitoring device that resides fully in the subcutaneous tissue has the potential to greatly improve the management of diabetes. Toward this goal, we have developed a competitive binding glucose sensing assay based on fluorescently labeled PEGylated concanavalin-A (PEGylated-TRITC-ConA) and mannotetraose (APTS-MT). In the present work, we sought to contain this assay within the hollow central cavity of a cylindrical hydrogel membrane, permitting eventual subcutaneous implantation and optical probing through the skin. A "self-cleaning" hydrogel was utilized because of its ability to cyclically deswell/reswell in vivo, which is expected to reduce biofouling and therefore extend the sensor lifetime. Thus, we prepared a hollow, cylindrical hydrogel based on a thermoresponsive electrostatic double network design composed of N-isopropylacrylamide and 2-acrylamido-2-methylpropanesulfonic acid. Next, a layer-by-layer (LbL) coating was applied to the inner wall of the central cavity of the cylindrical membrane. It consisted of 5, 10, 15, 30, or 40 alternating bilayers of positively charged poly(diallyldimethylammonium chloride) and negatively charged poly(sodium 4-styrenesulfonate). With 30 bilayers, the leaching of the smaller-sized component of the assay (APTS-MT) from the membrane cavity was substantially reduced. Moreover, this LbL coating maintained glucose diffusion across the hydrogel membrane. In terms of sensor functionality, the assay housed in the hydrogel membrane cavity tracked changes in glucose concentration (0 to 600 mg/dL) with a mean absolute relative difference of ∼11%.

Label-free spectrochemical probe for determination of hemoglobin glycation in clinical blood samples

Glycated hemoglobin, HbA1c, is an important biomarker that reveals the average value of blood glucose over the preceding 3 months. While significant recent attention has been focused on the use of optical and direct molecular spectroscopic methods for determination of HbA1c, a facile test that minimizes sample preparation needs and turnaround time still remains elusive. Here, we report a label-free approach for identifying low, mid and high-HbA1c groups in hemolysate and in whole blood samples featuring resonance Raman (RR) spectroscopy and support vector machine (SVM)-based classification of spectral patterns. The diagnostic power of RR measurements stems from its selective enhancement of hemoglobin-specific features, which simultaneously minimizes the blood matrix spectral interference and permits detection in the native solution. In this pilot study, our spectroscopic observations reveal that glycation of hemoglobin results in subtle but reproducible changes even when detected in the whole blood matrix. Leveraging SVM analysis of the principal component scores determined from the RR spectra, we show high degree of accuracy in classifying clinical specimen. We envisage that the promising findings will pave the way for more extensive clinical specimen investigations with the ultimate goal of translating molecular spectroscopy for routine point-of-care testing.

Critical-depth Raman spectroscopy enables home-use non-invasive glucose monitoring

One of the most ambitious endeavors in the field of diabetes technology is non-invasive glucose sensing. In the past decades, a number of different technologies have been assessed, but none of these have found its entry into general clinical use. We report on the development of a table-top confocal Raman spectrometer that was used in the home of patients with diabetes and operated for extended periods of time unsupervised and without recalibration. The system is based on measurement of glucose levels at a ‘critical depth’ in the skin, specifically in the interstitial fluid located below the stratum corneum but above the underlying adipose tissue layer. The region chosen for routine glucose measurements was the base of the thumb (the thenar). In a small clinical study, 35 patients with diabetes analyzed their interstitial fluid glucose for a period of 60 days using the new critical-depth Raman (CD-Raman) method and levels were correlated to reference capillary blood glucose values using a standard finger-stick and test strip product. The calibration of the CD-Raman system was stable for > 10 days. Measurement performance for glucose levels present at, or below, a depth of ~250μm below the skin surface was comparable to that reported for currently available invasive continuous glucose monitors. In summary, using the CD-Raman technology we have demonstrated the first successful use of a non-invasive glucose monitor in the home.

Noninvasive glucose monitoring using mid-infrared absorption spectroscopy based on a few wavenumbers

A method for performing noninvasive blood glucose measurements was developed. The method is based on mid-infrared absorption spectroscopy and uses only a few wavenumbers to measure blood glucose levels in vivo unconditionally. We found that the regression of blood glucose levels using only three wavenumbers, which were selected using a series cross-validation technique, realized accuracies comparable to those of cases in which a greater number of wavenumbers are used. In addition, we demonstrated the performance of this model through correlations among different types of data.

Advanced monitoring and control of pharmaceutical production processes with Pichia pastoris by using Raman spectroscopy and multivariate calibration methods

This contribution includes an investigation of the applicability of Raman spectroscopy as a PAT analyzer in cyclic production processes of a potential Malaria vaccine with Pichia pastoris. In a feasibility study, Partial Least Squares Regression (PLSR) models were created off-line for cell density and concentrations of glycerol, methanol, ammonia and total secreted protein. Relative cross validation errors RMSE_cvrel range from 2.87% (glycerol) to 11.0% (ammonia). In the following, on-line bioprocess monitoring was tested for cell density and glycerol concentration. By using the nonlinear Support Vector Regression (SVR) method instead of PLSR, the error RMSE_Prel for cell density was reduced from 5.01 to 2.94%. The high potential of Raman spectroscopy in combination with multivariate calibration methods was demonstrated by the implementation of a closed loop control for glycerol concentration using PLSR. The strong nonlinear behavior of exponentially increasing control disturbances was met with a feed-forward control and adaptive correction of control parameters. In general the control procedure works very well for low cell densities. Unfortunately, PLSR models for glycerol concentration are strongly influenced by a correlation with the cell density. This leads to a failure in substrate prediction, which in turn prevents substrate control at cell densities above 16 g/L.

Floating reference position-based correction method for near-infrared spectroscopy in long-term glucose concentration monitoring

We present a floating reference position (FRP)-based drift correction method for near-infrared (NIR) spectroscopy-based long-term blood glucose concentration (BGC) monitoring. Previously, we reported that it is difficult to quantify the systematic drift caused by the fluctuation of incident light intensity at different source–detector (SD) separations based on the absolute FRP change. We use the relative FRP change as a baseline reference to quantitatively characterize the signal drift at different SD separations. For the wavelengths that were used, a uniform equation was developed to describe the relationship between the drift and the relative FRP change. With the help of this equation, the correction can easily be performed by subtracting the systematic drift estimated by the equation. A theoretical analysis and an experimental phantom study demonstrated that our method could be used for systematic drift correction in NIR spectroscopy for long-term BGC monitoring. Moreover, the analysis method can also be referenced to reduce drifts from multiple sources.

Noninvasive Monitoring of Blood Glucose with Raman Spectroscopy

The successful development of a noninvasive blood glucose sensor that can operate reliably over sustained periods of time has been a much sought after but elusive goal in diabetes management. Since diabetes has no well-established cure, control of elevated glucose levels is critical for avoiding severe secondary health complications in multiple organs including the retina, kidney and vasculature. While fingerstick testing continues to be the mainstay of blood glucose detection, advances in electrochemical sensing-based minimally invasive approaches have opened the door for alternate methods that would considerably improve the quality of life for people with diabetes. In the quest for better sensing approaches, optical technologies have surfaced as attractive candidates as researchers have sought to exploit the endogenous contrast of glucose, notably its absorption, scattering, and polarization properties. Vibrational spectroscopy, especially spontaneous Raman scattering, has exhibited substantial promise due to its exquisite molecular specificity and minimal interference of water in the spectral profiles acquired from the blood-tissue matrix. Yet, it has hitherto been challenging to leverage the Raman scattering signatures of glucose for prediction in all but the most basic studies and under the least demanding conditions. In this Account, we discuss the newly developed array of methodologies that address the key challenges in measuring blood glucose accurately using Raman spectroscopy and unlock new prospects for translation to sustained noninvasive measurements in people with diabetes. Owing to the weak intensity of spontaneous Raman scattering, recent research has focused on enhancement of signals from the blood constituents by designing novel excitation-collection geometries and tissue modulation methods while our attempts have led to the incorporation of nonimaging optical elements. Additionally, invoking mass transfer modeling into chemometric algorithms has not only addressed the physiological lag between the actual blood glucose and the measured interstitial fluid glucose values but also offered a powerful tool for predictive measurements of hypoglycemia. This framework has recently been extended to provide longitudinal tracking of glucose concentration without necessitating extensive a priori concentration information. These findings are advanced by the results of recent glucose tolerance studies in human subjects, which also hint at the need for designing nonlinear calibration models that can account for subject-to-subject variations in skin heterogeneity and hematocrit levels. Together, the emerging evidence underscores the promise of a blood withdrawal-free optical platform-featuring a combination of high-throughput Raman spectroscopic instrumentation and data analysis of subtle variations in spectral expression-for diabetes screening in the clinic and, ultimately, for personalized monitoring.

From near-infrared and Raman to surface-enhanced Raman spectroscopy: progress, limitations and perspectives in bioanalysis

Over recent decades, spreading environmental concern entailed the expansion of green chemistry analytical tools. Vibrational spectroscopy, belonging to this class of analytical tool, is particularly interesting taking into account its numerous advantages such as fast data acquisition and no sample preparation. In this context, near-infrared, Raman and mainly surface-enhanced Raman spectroscopy (SERS) have thus gained interest in many fields including bioanalysis. The two former techniques only ensure the analysis of concentrated compounds in simple matrices, whereas the emergence of SERS improved the performances of vibrational spectroscopy to very sensitive and selective analyses. Complex SERS substrates were also developed enabling biomarker measurements, paving the way for SERS immunoassays. Therefore, in this paper, the strengths and weaknesses of these techniques will be highlighted with a focus on recent progress.

Intraoperative diagnosis of benign and malignant breast tissues by fourier transform infrared spectroscopy and support vector machine classification

BACKGROUND

Fourier transform infrared (FTIR) spectroscopy has shown its unique advantages in distinguishing cancerous tissue from normal one. The aim of this study was to establish a quick and accurate diagnostic method of FTIR spectroscopy to differentiate malignancies from benign breast tissues intraoperatively.

MATERIALS AND METHODS

In this study, a total of 100 breast tissue samples obtained from 100 patients were taken on surgery. All tissue samples were scanned for spectra intraoperatively before being processed for histopathological diagnosis. Standard normal variate (SNV) method was adopted to reduce scatter effects. Support vector machine (SVM) classification was used to discriminate spectra between malignant and benign breast tissues. Leave-one-out cross validation (LOOCV) was used to evaluate the discrimination.

RESULTS

According to histopathological examination, 50 cases were diagnosed as fibroadenoma and 50 cases as invasive ductal carcinoma. The results of SVM algorithm showed that the sensitivity, specificity and accuracy rate of this method are 90.0%, 98.0% and 94.0%, respectively.

CONCLUSIONS

FTIR spectroscopy technique in combination with SVM classification could be an accurate, rapid and objective tool to differentiate malignant from benign tumors during operation. Our studies establish the feasibility of FTIR spectroscopy with chemometrics method to guide surgeons during the surgery as an effective supplement for pathological diagnosis on frozen section.

Emerging trends in optical sensing of glycemic markers for diabetes monitoring

In the past decade, considerable attention has been focused on the measurement of glycemic markers, such as glycated hemoglobin and glycated albumin, that provide retrospective indices of average glucose levels in the bloodstream. While these biomarkers have been regularly used to monitor long-term glucose control in established diabetics, they have also gained traction in diabetic screening. Detection of such glycemic markers is challenging, especially in a point-of-care setting, due to the stringent requirements for sensitivity and robustness. A number of non-separation based measurement strategies were recently proposed, including photonic tools that are well suited to reagent-free marker quantitation. Here, we critically review these methods while focusing on vibrational spectroscopic methods, which offer highly specific molecular fingerprinting capability. We examine the underlying principles and the utility of these approaches as reagentless assays capable of multiplexed detection of glycemic markers and also the challenges in their eventual use in the clinic.

Calibration diagnostic and updating strategy based on quantitative modeling of near-infrared spectral residuals

Near-infrared spectral residuals are used to develop diagnostic and model updating procedures to enhance the performance of multivariate calibrations.

Noninvasive measurement of glucose in artificial plasma with near-infrared and Raman spectroscopy

The goal of this research was to develop a method for noninvasive blood glucose assay. Near-infrared (NIR) spectroscopy and Raman spectroscopy, two more promising techniques compared to other methods, were investigated in two kinds of artificial plasma (AP). Calibration models were generated by performing partial least squares (PLS) regression and optimized individually by considering spectral range, spectral pretreatment methods, and number of model factors. The two spectroscopic models were validated for the determination of glucose, and the results show that the two spectroscopic models established are robust, accurate, and repeatable. Compared to Raman spectroscopy, the performance of NIR spectroscopy was much better, with lower root mean square errors of cross-validation (RMSECV) of 0.128 and 0.094 mg/ml, lower root mean square errors of validation (RMSEP) of 0.061 and 0.046 mg/ml, higher correlation coefficients (R) of 99.15% and 99.55%, and higher residual predictive deviations (RPD) of 10.8 and 15.0 for artificial plasma I and II, respectively.

Assessment and correction of turbidity effects on Raman observations of chemicals in aqueous solutions

Improvements in diode laser, fiber optic, and data acquisition technologies are enabling increased use of Raman spectroscopic techniques for both in lab and in situ water analysis. Aqueous media encountered in the natural environment often contain suspended solids that can interfere with spectroscopic measurements, yet removal of these solids, for example, via filtration, can have even greater adverse effects on the extent to which subsequent measurements are representative of actual field conditions. In this context, this study focuses on evaluation of turbidity effects on Raman spectroscopic measurements of two common environmental pollutants in aqueous solution: ammonium nitrate and trichloroethylene. The former is typically encountered in the runoff from agricultural operations and is a strong scatterer that has no significant influence on the Raman spectrum of water. The latter is a commonly encountered pollutant at contaminated sites associated with degreasing and cleaning operations and is a weak scatterer that has a significant influence on the Raman spectrum of water. Raman observations of each compound in aqueous solutions of varying turbidity created by doping samples with silica flour with grain sizes ranging from 1.6 to 5.0 μm were employed to develop relationships between observed Raman signal strength and turbidity level. Shared characteristics of these relationships were then employed to define generalized correction methods for the effect of turbidity on Raman observations of compounds in aqueous solution.

Raman spectroscopy: in vivo quick response code of skin physiological status

Dermatologists need to combine different clinically relevant characteristics for a better understanding of skin health. These characteristics are usually measured by different techniques, and some of them are highly time consuming. Therefore, a predicting model based on Raman spectroscopy and partial least square (PLS) regression was developed as a rapid multiparametric method. The Raman spectra collected from the five uppermost micrometers of 11 healthy volunteers were fitted to different skin characteristics measured by independent appropriate methods (transepidermal water loss, hydration, pH, relative amount of ceramides, fatty acids, and cholesterol). For each parameter, the obtained PLS model presented correlation coefficients higher than R2=0.9. This model enables us to obtain all the aforementioned parameters directly from the unique Raman signature. In addition to that, in-depth Raman analyses down to 20 μm showed different balances between partially bound water and unbound water with depth. In parallel, the increase of depth was followed by an unfolding process of the proteins. The combinations of all these information led to a multiparametric investigation, which better characterizes the skin status. Raman signal can thus be used as a quick response code (QR code). This could help dermatologic diagnosis of physiological variations and presents a possible extension to pathological characterization.

Development and comparative assessment of Raman spectroscopic classification algorithms for lesion discrimination in stereotactic breast biopsies with microcalcifications

Microcalcifications are an early mammographic sign of breast cancer and a target for stereotactic breast needle biopsy. Here, we develop and compare different approaches for developing Raman classification algorithms to diagnose invasive and in situ breast cancer, fibrocystic change and fibroadenoma that can be associated with microcalcifications. In this study, Raman spectra were acquired from tissue cores obtained from fresh breast biopsies and analyzed using a constituent-based breast model. Diagnostic algorithms based on the breast model fit coefficients were devised using logistic regression, C4.5 decision tree classification, k-nearest neighbor (k -NN) and support vector machine (SVM) analysis, and subjected to leave-one-out cross validation. The best performing algorithm was based on SVM analysis (with radial basis function), which yielded a positive predictive value of 100% and negative predictive value of 96% for cancer diagnosis. Importantly, these results demonstrate that Raman spectroscopy provides adequate diagnostic information for lesion discrimination even in the presence of microcalcifications, which to the best of our knowledge has not been previously reported.

A hybrid method to recognize 3D object

A hybrid method using the support vector machine (SVM) correlation filter and the phase-shift interferometry (PSI) holography is proposed to recognize 3D object, which can improve the correct decision rate and resist the distortion of object rotation and noise. The different images of two types of both in-plane and out-of-plane rotated object recorded by digital holography are reconstructed. The reconstructed images of two types are selected to synthesize the SVM correlation filter, respectively. To compare the correct decision rates of the SVM correlation filter with other three ones, it is found that the experimental result is better in rotation resistance and noise tolerance.

Noninvasive glucose detection in human skin using wavelength modulated differential laser photothermal radiometry

Noninvasive glucose monitoring will greatly improve diabetes management. We applied Wavelength-Modulated Differential Laser Photothermal Radiometry (WM-DPTR) to noninvasive glucose measurements in human skin in vitro in the mid-infrared range. Glucose measurements in human blood serum diffused into a human skin sample (1 mm thickness from abdomen) in the physiological range (21-400 mg/dl) demonstrated high sensitivity and accuracy to meet wide clinical detection requirements. It was found that the glucose sensitivity could be tuned by adjusting the intensity ratio and phase difference of the two laser beams in the WM-DPTR system. The measurement results demonstrated the feasibility of the development of WM-DPTR into a clinically viable noninvasive glucose biosensor.

Investigation of noise-induced instabilities in quantitative biological spectroscopy and its implications for noninvasive glucose monitoring

Over the past decade, optical spectroscopy has been employed in combination with multivariate chemometric models to investigate a wide variety of diseases and pathological conditions, primarily due to its excellent chemical specificity and lack of sample preparation requirements. Despite promising results in several proof-of-concept studies, its translation to the clinical setting has often been hindered by inadequate accuracy of the conventional spectroscopic models. To address this issue and the possibility of curved (nonlinear) effects in the relationship between the concentrations of the analyte of interest and the mixture spectra (due to fluctuations in sample and environmental conditions), support vector machine-based least-squares nonlinear regression (LS-SVR) has been recently proposed. In this paper, we investigate the robustness of this methodology to noise-induced instabilities and present an analytical formula for estimating modeling precision as a function of measurement noise and model parameters. This formalism can be readily used to evaluate uncertainty in information extracted from spectroscopic measurements, particularly important for rapid-acquisition biomedical applications. Subsequently, using field data (Raman spectra) acquired from a glucose clamping study on an animal model subject, we perform the first systematic investigation of the relative effect of additive interference components (namely, noise in prediction spectra, calibration spectra, and calibration concentrations) on the prediction error of nonlinear spectroscopic models. Our results show that the LS-SVR method gives more accurate results and is substantially more robust to additive noise when compared with conventional regression methods such as partial least-squares regression (PLS), when careful selection of the LS-SVR model parameters are performed. We anticipate that these results will be useful for uncertainty estimation in similar biomedical applications where the precision of measurements and its response to noise in the data set is as important, if not more so, than the generic accuracy level.

Logic minimization and rule extraction for identification of functional sites in molecular sequences

Background

Logic minimization is the application of algebraic axioms to a binary dataset with the purpose of reducing the number of digital variables and/or rules needed to express it. Although logic minimization techniques have been applied to bioinformatics datasets before, they have not been used in classification and rule discovery problems. In this paper, we propose a method based on logic minimization to extract predictive rules for two bioinformatics problems involving the identification of functional sites in molecular sequences: transcription factor binding sites (TFBS) in DNA and O-glycosylation sites in proteins. TFBS are important in various developmental processes and glycosylation is a posttranslational modification critical to protein functions.

Methods

In the present study, we first transformed the original biological dataset into a suitable binary form. Logic minimization was then applied to generate sets of simple rules to describe the transformed dataset. These rules were used to predict TFBS and O-glycosylation sites. The TFBS dataset is obtained from the TRANSFAC database, while the glycosylation dataset was compiled using information from OGLYCBASE and the Swiss-Prot Database.

We performed the same predictions using two standard classification techniques, Artificial Neural Networks (ANN) and Support Vector Machines (SVM), and used their sensitivities and positive predictive values as benchmarks for the performance of our proposed algorithm. SVM were also used to reduce the number of variables included in the logic minimization approach.

Results

For both TFBS and O-glycosylation sites, the prediction performance of the proposed logic minimization method was generally comparable and, in some cases, superior to the standard ANN and SVM classification methods with the advantage of providing intelligible rules to describe the datasets. In TFBS prediction, logic minimization produced a very small set of simple rules. In glycosylation site prediction, the rules produced were also interpretable and the most popular rules generated appeared to correlate well with recently reported hydrophilic/hydrophobic enhancement values of amino acids around possible O-glycosylation sites. Experiments with Self-Organizing Neural Networks corroborate the practical worth of the logic minimization method for these case studies.

Conclusions

The proposed logic minimization algorithm provides sets of rules that can be used to predict TFBS and O-glycosylation sites with sensitivity and positive predictive value comparable to those from ANN and SVM. Moreover, the logic minimization method has the additional capability of generating interpretable rules that allow biological scientists to correlate the predictions with other experimental results and to form new hypotheses for further investigation. Additional experiments with alternative rule-extraction techniques demonstrate that the logic minimization method is able to produce accurate rules from datasets with large numbers of variables and limited numbers of positive examples.

Incorporation of support vector machines in the LIBS toolbox for sensitive and robust classification amidst unexpected sample and system variability

Despite the intrinsic elemental analysis capability and lack of sample preparation requirements, laser-induced breakdown spectroscopy (LIBS) has not been extensively used for real-world applications, e.g., quality assurance and process monitoring. Specifically, variability in sample, system, and experimental parameters in LIBS studies present a substantive hurdle for robust classification, even when standard multivariate chemometric techniques are used for analysis. Considering pharmaceutical sample investigation as an example, we propose the use of support vector machines (SVM) as a nonlinear classification method over conventional linear techniques such as soft independent modeling of class analogy (SIMCA) and partial least-squares discriminant analysis (PLS-DA) for discrimination based on LIBS measurements. Using over-the-counter pharmaceutical samples, we demonstrate that the application of SVM enables statistically significant improvements in prospective classification accuracy (sensitivity), because of its ability to address variability in LIBS sample ablation and plasma self-absorption behavior. Furthermore, our results reveal that SVM provides nearly 10% improvement in correct allocation rate and a concomitant reduction in misclassification rates of 75% (cf. PLS-DA) and 80% (cf. SIMCA)-when measurements from samples not included in the training set are incorporated in the test data-highlighting its robustness. While further studies on a wider matrix of sample types performed using different LIBS systems is needed to fully characterize the capability of SVM to provide superior predictions, we anticipate that the improved sensitivity and robustness observed here will facilitate application of the proposed LIBS-SVM toolbox for screening drugs and detecting counterfeit samples, as well as in related areas of forensic and biological sample analysis.

Raman spectroscopy-based sensitive and specific detection of glycated hemoglobin

In recent years, glycated hemoglobin (HbA1c) has been increasingly accepted as a functional metric of mean blood glucose in the treatment of diabetic patients. Importantly, HbA1c provides an alternate measure of total glycemic exposure due to the representation of blood glucose throughout the day, including post-prandially. In this article, we propose and demonstrate the potential of Raman spectroscopy as a novel analytical method for quantitative detection of HbA1c, without using external dyes or reagents. Using the drop coating deposition Raman (DCDR) technique, we observe that the nonenzymatic glycosylation (glycation) of the hemoglobin molecule results in subtle but discernible and highly reproducible changes in the acquired spectra, which enable the accurate determination of glycated and nonglycated hemoglobin using standard chemometric methods. The acquired Raman spectra display excellent reproducibility of spectral characteristics at different locations in the drop and show a linear dependence of the spectral intensity on the analyte concentration. Furthermore, in hemolysate models, the developed multivariate calibration models for HbA1c show a high degree of prediction accuracy and precision--with a limit of detection that is a factor of ~15 smaller than the lowest physiological concentrations encountered in clinical practice. The excellent accuracy and reproducibility achieved in this proof-of-concept study opens substantive avenues for characterization and quantification of the glycosylation status of (therapeutic) proteins, which are widely used for biopharmaceutical development. We also envision that the proposed approach can provide a powerful tool for high-throughput HbA1c sensing in multicomponent mixtures and potentially in hemolysate and whole blood lysate samples.

Raman spectroscopy provides a powerful diagnostic tool for accurate determination of albumin glycation

We present the first demonstration of glycated albumin detection and quantification using Raman spectroscopy without the addition of reagents. Glycated albumin is an important marker for monitoring the long-term glycemic history of diabetics, especially as its concentrations, in contrast to glycated hemoglobin levels, are unaffected by changes in erythrocyte life times. Clinically, glycated albumin concentrations show a strong correlation with the development of serious diabetes complications including nephropathy and retinopathy. In this article, we propose and evaluate the efficacy of Raman spectroscopy for determination of this important analyte. By utilizing the pre-concentration obtained through drop-coating deposition, we show that glycation of albumin leads to subtle, but consistent, changes in vibrational features, which with the help of multivariate classification techniques can be used to discriminate glycated albumin from the unglycated variant with 100% accuracy. Moreover, we demonstrate that the calibration model developed on the glycated albumin spectral dataset shows high predictive power, even at substantially lower concentrations than those typically encountered in clinical practice. In fact, the limit of detection for glycated albumin measurements is calculated to be approximately four times lower than its minimum physiological concentration. Importantly, in relation to the existing detection methods for glycated albumin, the proposed method is also completely reagent-free, requires barely any sample preparation and has the potential for simultaneous determination of glycated hemoglobin levels as well. Given these key advantages, we believe that the proposed approach can provide a uniquely powerful tool for quantification of glycation status of proteins in biopharmaceutical development as well as for glycemic marker determination in routine clinical diagnostics in the future.

A novel non-imaging optics based Raman spectroscopy device for transdermal blood analyte measurement

Due to its high chemical specificity, Raman spectroscopy has been considered to be a promising technique for non-invasive disease diagnosis. However, during Raman excitation, less than one out of a million photons undergo spontaneous Raman scattering and such weakness in Raman scattered light often require highly efficient collection of Raman scattered light for the analysis of biological tissues. We present a novel non-imaging optics based portable Raman spectroscopy instrument designed for enhanced light collection. While the instrument was demonstrated on transdermal blood glucose measurement, it can also be used for detection of other clinically relevant blood analytes such as creatinine, urea and cholesterol, as well as other tissue diagnosis applications. For enhanced light collection, a non-imaging optical element called compound hyperbolic concentrator (CHC) converts the wide angular range of scattered photons (numerical aperture (NA) of 1.0) from the tissue into a limited range of angles accommodated by the acceptance angles of the collection system (e.g., an optical fiber with NA of 0.22). A CHC enables collimation of scattered light directions to within extremely narrow range of angles while also maintaining practical physical dimensions. Such a design allows for the development of a very efficient and compact spectroscopy system for analyzing highly scattering biological tissues. Using the CHC-based portable Raman instrument in a clinical research setting, we demonstrate successful transdermal blood glucose predictions in human subjects undergoing oral glucose tolerance tests.

Wavelength selection-based nonlinear calibration for transcutaneous blood glucose sensing using Raman spectroscopy

While Raman spectroscopy provides a powerful tool for noninvasive and real time diagnostics of biological samples, its translation to the clinical setting has been impeded by the lack of robustness of spectroscopic calibration models and the size and cumbersome nature of conventional laboratory Raman systems. Linear multivariate calibration models employing full spectrum analysis are often misled by spurious correlations, such as system drift and covariations among constituents. In addition, such calibration schemes are prone to overfitting, especially in the presence of external interferences that may create nonlinearities in the spectra-concentration relationship. To address both of these issues we incorporate residue error plot-based wavelength selection and nonlinear support vector regression (SVR). Wavelength selection is used to eliminate uninformative regions of the spectrum, while SVR is used to model the curved effects such as those created by tissue turbidity and temperature fluctuations. Using glucose detection in tissue phantoms as a representative example, we show that even a substantial reduction in the number of wavelengths analyzed using SVR lead to calibration models of equivalent prediction accuracy as linear full spectrum analysis. Further, with clinical datasets obtained from human subject studies, we also demonstrate the prospective applicability of the selected wavelength subsets without sacrificing prediction accuracy, which has extensive implications for calibration maintenance and transfer. Additionally, such wavelength selection could substantially reduce the collection time of serial Raman acquisition systems. Given the reduced footprint of serial Raman systems in relation to conventional dispersive Raman spectrometers, we anticipate that the incorporation of wavelength selection in such hardware designs will enhance the possibility of miniaturized clinical systems for disease diagnosis in the near future.

Investigation of the specificity of Raman spectroscopy in non-invasive blood glucose measurements

Although several in vivo blood glucose measurement studies have been performed by different research groups using near-infrared (NIR) absorption and Raman spectroscopic techniques, prospective prediction has proven to be a challenging problem. An important issue in this case is the demonstration of causality of glucose concentration to the spectral information, especially as the intrinsic glucose signal is smaller compared with that of the other analytes in the blood-tissue matrix. Furthermore, time-dependent physiological processes make the relation between glucose concentration and spectral data more complex. In this article, chance correlations in Raman spectroscopy-based calibration model for glucose measurements are investigated for both in vitro (physical tissue models) and in vivo (animal model and human subject) cases. Different spurious glucose concentration profiles are assigned to the Raman spectra acquired from physical tissue models, where the glucose concentration is intentionally held constant. Analogous concentration profiles, in addition to the true concentration profile, are also assigned to the datasets acquired from an animal model during a glucose clamping study as well as a human subject during an oral glucose tolerance test. We demonstrate that the spurious concentration profile-based calibration models are unable to provide prospective predictions, in contrast to those based on actual concentration profiles, especially for the physical tissue models. We also show that chance correlations incorporated by the calibration models are significantly less in Raman as compared to NIR absorption spectroscopy, even for the in vivo studies. Finally, our results suggest that the incorporation of chance correlations for in vivo cases can be largely attributed to the uncontrolled physiological sources of variations. Such uncontrolled physiological variations could either be intrinsic to the subject or stem from changes in the measurement conditions.

Encapsulation of a Concanavalin A/dendrimer glucose sensing assay within microporated poly (ethylene glycol) microspheres

Proper management of diabetes requires the frequent measurement of a patient's blood glucose level. To create a long-term, minimally-invasive sensor that is sensitive to physiological concentrations of glucose a fluorescent glucose sensing assay using a competitive binding approach between fluorescently tagged Concanavalin-A (Con-A) and glycodendrimer is being developed. Until now, the essential step of effectively encapsulating this aggregative sensing assay while allowing a reversible response has yet to be reported. In this paper, a microporation technique is described in which microspheres are synthesized in a manner that creates fluid-filled pores within a poly (ethylene glycol) hydrogel. This dual-nature technique creates hydrophilic, biocompatible microcapsules in which the aggregative binding kinetics of the sensing assay within the pores are not constrained by spatial fixation in the hydrogel matrix. Confocal images displaying the localization of pockets filled with the assay within the polymeric matrix are presented in this paper. In addition, fluorescent responses to varying glucose concentrations, leaching studies, and long-term functionality of the encapsulated assay are demonstrated. To our knowledge, this is the first time that the Con-A/glycodendrimer assay has been shown to be reversible and repeatable within hydrogel spheres, including the display of functionality up to fourteen days under ambient conditions.

Rapid and accurate determination of tissue optical properties using least-squares support vector machines

Diffuse reflectance spectroscopy (DRS) has been extensively applied for the characterization of biological tissue, especially for dysplasia and cancer detection, by determination of the tissue optical properties. A major challenge in performing routine clinical diagnosis lies in the extraction of the relevant parameters, especially at high absorption levels typically observed in cancerous tissue. Here, we present a new least-squares support vector machine (LS-SVM) based regression algorithm for rapid and accurate determination of the absorption and scattering properties. Using physical tissue models, we demonstrate that the proposed method can be implemented more than two orders of magnitude faster than the state-of-the-art approaches while providing better prediction accuracy. Our results show that the proposed regression method has great potential for clinical applications including in tissue scanners for cancer margin assessment, where rapid quantification of optical properties is critical to the performance.