Raman spectroscopy combined with a support vector machine for differentiating between feeding male and female infants mother's milk

What they forgot to tell you about machine learning with an application to pharmaceutical manufacturing

Predictive models (a.k.a. machine learning models) are ubiquitous in all stages of drug research, safety, development, manufacturing, and marketing. The results of these models are used inside and outside of pharmaceutical companies for the purpose of understanding scientific processes and for predicting characteristics of new samples or patients. While there are many resources that describe such models, there are few that explain how to develop a robust model that extracts the highest possible performance from the available data, especially in support of pharmaceutical applications. This tutorial will describe pitfalls and best practices for developing and validating predictive models with a specific application to a monitoring a pharmaceutical manufacturing process. The pitfalls and best practices will be highlighted to call attention to specific points that are not generally discussed in other resources.

First Trimester Prediction of Preterm Birth in Patient Plasma with Machine-Learning-Guided Raman Spectroscopy and Metabolomics

Preterm birth (PTB) is the leading cause of infant deaths globally. Current clinical measures often fail to identify women who may deliver preterm. Therefore, accurate screening tools are imperative for early prediction of PTB. Here, we show that Raman spectroscopy is a promising tool for studying biological interfaces, and we examine differences in the maternal metabolome of the first trimester plasma of PTB patients and those that delivered at term (healthy). We identified fifteen statistically significant metabolites that are predictive of the onset of PTB. Mass spectrometry metabolomics validates the Raman findings identifying key metabolic pathways that are enriched in PTB. We also show that patient clinical information alone and protein quantification of standard inflammatory cytokines both fail to identify PTB patients. We show for the first time that synergistic integration of Raman and clinical data guided with machine learning results in an unprecedented 85.1% accuracy of risk stratification of PTB in the first trimester that is currently not possible clinically. Correlations between metabolites and clinical features highlight the body mass index and maternal age as contributors of metabolic rewiring. Our findings show that Raman spectral screening may complement current prenatal care for early prediction of PTB, and our approach can be translated to other patient-specific biological interfaces.

Accurate identification of living Bacillus spores using laser tweezers Raman spectroscopy and deep learning

Accurately, rapidly, and noninvasively identifying Bacillus spores can greatly contribute to controlling a plenty of infectious diseases. Laser tweezers Raman spectroscopy (LTRS) has confirmed to be a powerful tool for studying Bacillus spores at a single cell level. In this study, we constructed a single-cell Raman spectra dataset of living Bacillus spores and utilized deep learning approach to accurately, nondestructively identify Bacillus spores. The trained convolutional neural network (CNN) could efficiently extract tiny Raman spectra features of five spore species, and provide a prediction accuracy of specie identification as high as 100 %. Moreover, the spectral feature differences in three Raman bands at 660, 826, and 1017 cm^-1 were confirmed to mostly contribute to producing such high prediction accuracy. In addition, optimal CNN model was employed to monitor and identify sporulation process at different metabolic phases in one growth cycle. The obtained average prediction accuracy of metabolic phase identification was approximately 88 %. It can be foreseen that, LTRS combined with CNN approach have great potential for accurately identifying spore species and metabolic phases at a single cell level, and can be gradually extended to perform identification for many unculturable bacteria growing in soil, water, and food.

Fatty Acid Determination in Human Milk Using Attenuated Total Reflection Infrared Spectroscopy and Solvent-Free Lipid Separation

This study introduces the first mid-infrared (IR)-based method for determining the fatty acid composition of human milk. A representative milk lipid fraction was obtained by applying a rapid and solvent-free two-step centrifugation method. Attenuated total reflection Fourier transform infrared (ATR FT-IR) spectroscopy was applied to record absorbance spectra of pure milk fat. The obtained spectra were compared to whole human milk transmission spectra, revealing the significantly higher degree of fatty acid-related spectral features in ATR FT-IR spectra. Partial least squares (PLS)-based multivariate regression equations were established by relating ATR FT-IR spectra to fatty acid reference concentrations, obtained with gas chromatography-mass spectrometry (GC-MS). Good predictions were achieved for the most important fatty acid sum parameters: saturated fatty acids (SAT, R²_CV = 0.94), monounsaturated fatty acids (MONO, R²_CV = 0.85), polyunsaturated fatty acids (PUFA, R²_CV = 0.87), unsaturated fatty acids (UNSAT, R²_CV = 0.91), short-chain fatty acids (SCFA, R²_CV = 0.79), medium-chain fatty acids (MCFA, R²_CV = 0.97), and long-chain fatty acids (LCFA, R²_CV = 0.88). The PLS selectivity ratio (SR) was calculated in order to optimize and verify each individual calibration model. All mid-IR regions with high SR could be assigned to absorbances from fatty acids, indicating high validity of the obtained models.

Evaluation of the changes in human milk lipid composition and conformational state with Raman spectroscopy during a breastfeed

Human milk fat forms the main energy source for breastfed infants, and is highly variable in terms of concentration and composition. Understanding the changes in human milk lipid composition and conformational state during a breastfeed can provide insight into lipid synthesis and secretion in the mammary gland. Therefore, the aim of this study was to evaluate human milk fatty acid length, degree of unsaturation (lipid composition) and lipid phase (lipid conformational state) at different stages during a single breastfeed (fore-, bulk- and hindmilk). A total of 48 samples from 16 lactating subjects were investigated with confocal Raman spectroscopy. We did not observe any significant changes in lipid composition between fore-, bulk and hindmilk. A new finding from this study is that lipid conformational state at room temperature changed significantly during a breastfeed, from almost crystalline to almost liquid. This observation suggests that lipid synthesis in the mammary gland changes during a single breastfeed.

Qualitative and quantitative vibrational spectroscopic analysis of macronutrients in breast milk

Raman and attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy were used to analyze 208 breast milk samples as part of a larger research study. Comprehensive qualitative and quantitative analysis was carried out using chemometric methods: principal component analysis (PCA) and partial least squares (PLS) regression. The obtained information on the main macronutrients (protein, fat and carbohydrate) were primarily evaluated in relation to the available metadata of the samples, where study location and respective primary food sources revealed a stronger differentiation in fat composition than its absolute content. The limitations and challenges of using both spectroscopic techniques for the type of analysis are also highlighted.

Label-free detection of multiple genitourinary cancers from urine by surface-enhanced Raman spectroscopy

Detecting cancers through testing biological fluids, namely, "liquid biopsy", is noninvasive and shows great promise in cancer diagnosis, surveillance and screening. Many metabolites that may reflect cancer specificity are concentrated in and excreted through urine. In this study, urine samples were collected from healthy subjects and patients with bladder or prostate cancer. By using surface-enhanced Raman spectroscopy (SERS) with silver nanoparticles, urine sample spectra from 500-1800 cm^-1 were obtained. The spectra were classified by principal component analysis and linear discriminant analysis (PCA-LDA). The results showed that the classification accuracy of the model for healthy individuals, bladder cancer patients and prostate cancer patients was 91.9%, and the classification accuracy of the test set was 89%, which indicated that SERS combined with the PCA-LDA diagnostic algorithm could be used as a classification and diagnostic tool to detect and distinguish bladder cancer and prostate cancer through testing urine.

Comparison of Chemometric Problems in Food Analysis Using Non-Linear Methods

Food analysis is a challenging analytical problem, often addressed using sophisticated laboratory methods that produce large data sets. Linear and non-linear multivariate methods can be used to process these types of datasets and to answer questions such as whether product origin is accurately labeled or whether a product is safe to eat. In this review, we present the application of non-linear methods such as artificial neural networks, support vector machines, self-organizing maps, and multi-layer artificial neural networks in the field of chemometrics related to food analysis. We discuss criteria to determine when non-linear methods are better suited for use instead of traditional methods. The principles of algorithms are described, and examples are presented for solving the problems of exploratory analysis, classification, and prediction.

Dependency of the optical scattering properties of human milk on casein content and common sample preparation methods

SIGNIFICANCE

Quantifying human milk composition is important for daily nutritional management in neonatal intensive cares worldwide. Photonic solutions based on visible light can potentially aid in this analysis, as energy content of human milk depends largely on fat content, and the optical scattering properties of human milk predominantly depend on the size and concentration of fat globules. However, it is expected that human milk scattering changes upon homogenization, routinely done before analysis, which may affect fat globule size.

AIM

The first aim of this study was to investigate how the most common homogenization methods (gently inverting by hand, vortexing, and sonication) affect the optical properties of human milk. The second aim was to estimate the scattering contribution of casein micelles, the second most dominant scatterers in human milk.

APPROACH

We combined diffuse reflectance spectroscopy with spectroscopic optical coherence tomography to measure the scattering coefficient μs, reduced scattering coefficient μs', and anisotropy g between 450 and 600 nm.

RESULTS

Sonication induced the strongest changes in μs, μs', and g compared to the gently inverted samples (203%, 202%, and 7%, respectively, at 550 nm), but also vortexing changed μs' with 20%. Although casein micelles only showed a modest contribution to μs and g at 550 nm (7% and 1%, respectively), their contribution to μs' was 29%.

CONCLUSIONS

The scattering properties of human milk strongly depend on the homogenization method that is employed, and gentle inversion should be the preferred method. The contribution of casein micelles was relatively small for μs and g but considerably larger for μs'.

Determination of physiological lactate and pH by Raman spectroscopy

Lactate and pH values in human blood are important physiological parameters that can reflect serious pathological states as sepsis or hypoxia. In this study we demonstrate that we can determine lactate and pH values from blood samples by Raman spectroscopy combined with multivariate analysis at medically relevant resolution and reliability. The method prepares the base for new real-time in vivo analytics for a number of pathological cases and physiological control in competitive sports. We demonstrate the capability to resolve pH variations of 0.04 and lactate concentrations of 0.20 mM ex vivo.

A comparative study of machine learning classifiers for risk prediction of asthma disease

Asthma is a chronic disease characterized by wheezing, chest tightening and difficulty in breathing due to inflammation of lung airways. Early risk prediction of asthma is crucial for proper and effective management. This study presents the use of machine learning approach for risk prediction of asthma by evaluating Raman spectral variations between asthmatic as well as healthy sera samples. Specifically, Raman spectra from 150 asthma and 52 healthy control blood sera samples were acquired. Spectral analyses illustrated significant spectral variations (p < 0.0001) in the asthmatic samples when compared with healthy sera. The existing spectral differences were further exploited by using artificial neural network (ANN) along with support vector machine (SVM) and random forest (RF) algorithms towards machine-assisted classification of the two groups. Quantitative comparison of the evaluation metrics of the classification algorithms showed superior performance of SVM model. Our results indicate that Raman spectroscopy in tandem with SVM can be used in the diagnosis and machine-assisted classification of asthma patients with promising accuracy.

Optical properties of human milk

With human milk being the most important source of infant nutrition, the protection and support of breastfeeding are essential from a global health perspective. Nevertheless, relatively few objective methods are available to investigate human milk composition and lactation physiology when a mother experiences breastfeeding problems. We argue that optics and photonics offer promising opportunities for this purpose. Any research activity within this new application field starts with a thorough understanding on how light interacts with human milk. Therefore, the aim of this study was to investigate the full set of optical properties for human milk and the biological variability therein. Using a novel approach that combines spatially resolved diffuse reflectance spectroscopy (SR-DRS) and spectroscopic optical coherence tomography (sOCT) between 450 and 650 nm, we quantified the absorption coefficient µa , scattering coefficient µs , reduced scattering coefficient µs', anisotropy g and backscattering coefficient µb,NA of mature human milk from 14 participants released at different stages during a breastfeed (foremilk, bulk milk and hindmilk). Significant correlations were found between µa , µs , µs' and µb,NA and the biochemically determined fat concentration per sample (Rs = 0.38, Rs = 0.77, Rs = 0.80, Rs = 0.44 respectively). We explained the observed variations in the optical properties of human milk using Mie theory and the biological variability in both the concentration and size distribution of milk fat globules. In conclusion, we have provided a full set of optical properties for human milk, which can hopefully serve as a starting point for future biophotonic studies on human milk and the milk containing lactating breast.

Rapid and non-invasive screening of high renin hypertension using Raman spectroscopy and different classification algorithms

This study presents a rapid and non-invasive method to screen high renin hypertension using serum Raman spectroscopy combined with different classification algorithms. The serum samples taken from 24 high renin hypertension patients and 22 non-high renin hypertension samples were measured in this experiment. Tentative assignments of the Raman peaks in the measured serum spectra suggested specific biomolecular changes between the groups. Principal component analysis (PCA) was first used for feature extraction and reduced the dimension of high-dimension spectral data. Then, support vector machine (SVM), linear discriminant analysis (LDA) and k-nearest neighbor (KNN) algorithms were employed to establish the discriminant diagnostic models. The accuracies of 93.5%, 93.5% and 89.1% were obtained from PCA-SVM, PCA-LDA and PCA-KNN models, respectively. The results from our study demonstrate that the serum Raman spectroscopy technique combined with multivariate statistical methods have great potential for the screening of high renin hypertension. This technique could be used to develop a portable, rapid, and non-invasive device for screening high renin hypertension.

In vitro discrimination and classification of Microbial Flora of Poultry using two dispersive Raman spectrometers (microscope and Portable Fiber-Optic systems) in tandem with chemometric analysis

Discrimination and classification of eight strains related to meat spoilage and pathogenic microorganisms commonly found in poultry meat were successfully carried out using two dispersive Raman spectrometers (Microscope and Portable Fiber-Optic systems) in combination with chemometric methods. Principal components analysis (PCA) and multi-class support vector machines (MC-SVM) were applied to develop discrimination and classification models. These models were certified using validation data sets which were successfully assigned to the correct bacterial species and even to the right strain. The discrimination of bacteria down to the strain level was performed for the pre-processed spectral data using a 3-stage model based on PCA. The spectral features and differences among the species on which the discrimination was based were clarified through PCA loadings. In MC-SVM the pre-processed spectral data was subjected to PCA and utilized to build a classification model. When using the first two components, the accuracy of the MC-SVM model was 97.64% and 93.23% for the validation data collected by the Raman Microscope and the Portable Fiber-Optic Raman system, respectively. The accuracy reached 100% for the validation data by using the first eight and ten PC's from the data collected by Raman Microscope and by Portable Fiber-Optic Raman system, respectively. The results reflect the strong discriminative power and the high performance of the developed models, the suitability of the pre-processing method used in this study and that the low accuracy of the Portable Fiber-Optic Raman system does not adversely affect the discriminative power of the developed models.

Demonstrating the application of Raman spectroscopy together with chemometric technique for screening of asthma disease

Medical biophotonic tools provide new sources of diagnostic information regarding the state of human health that are used in managing patient care. In our current study, Raman spectroscopy, together with the chemometric technique, has successfully been demonstrated for the screening of asthma disease. Raman spectra of sera samples from asthmatic patients as well as healthy (control) volunteers have been recorded at 532 nm excitation. In healthy sera, three highly reproducible Raman peaks assigned to β-carotene have been detected. Their sensitive detection is facilitated due to the resonance Raman effect. In contrast, in asthmatic patients sera, the peaks assigned to β-carotene are either diminished or suppressed accompanied by other new Raman peaks. These new peaks most probably arise due to an elevated level of proteins, which could be used to identify/differentiate between asthma and non-asthma samples. Furthermore, a partial least squares discrimination analysis (PLS-DA) model was developed and applied on the Raman spectra of diseased as well as healthy samples, which successfully classified them. The correlation coefficient (r2) of the model was determined as 0.965. Similarly, the root mean square errors in cross-validation (RMSECV) and in the prediction (RMSECP) are 0.09 and 0.25, respectively. PLS-DA has the potential to be incorporated in a microcontroller's code attached with a hand-held Raman spectrometer for screening purposes in asthma, which is a disease of great concern for the clinicians, especially in children.