No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Baseline Correction with Asymmetric Least Squares Smoothing
Abstract
Most baseline problems in instrumental methods are characterized by a smooth baseline and a superimposed signal that carries the analytical information: a se-ries of peaks that are either all positive or all negative. We combine a smoother with asymmetric weighting of deviations from the (smooth) trend get an effective baseline estimator. It is easy to use, fast and keeps the analytical peak signal in-tact. No prior information about peak shapes or baseline (polynomial) is needed by the method. The performance is illustrated by simulation and applications to real data.
... Hence, PCA is first used to identify the variability region in the FTIR spectra responsible for the inter and intra-molecular difference between the RS samples. Also, this study proposes a fast and straightforward variable selection method by identifying the relevant peaks using a difference spectrum, a modified FTIR spectrum constructed using the asymmetric least squares smoothing (AsLS) algorithm [28][29][30]. Investigations into the molecular vibrations responsible for the identified peaks revealed that these peaks were associated with cellulose, hemicellulose and lignin. Interestingly, PCA using these peaks efficiently categorized the biomass based on their compositional similarity and dissimilarity. ...
... First, an approach similar to Sim et al. [35] has been employed for de-noising the processed spectra. Smoothing of these spectra was then carried out using the asymmetric least squares smoothing (AsLS) method based on the Whittaker smoother [28][29][30]. For processed spectra à i , consider ŝ i as the smooth estimator, which was smooth and dependent on à i . ...
... For processed spectra à i , consider ŝ i as the smooth estimator, which was smooth and dependent on à i . The AsLS algorithm for the determination of weights, as detailed by Eilers et al. [28] is by minimizing the penalized least squares function (Eq. 1) ...
The efficiency of lignocellulosic biomass conversion to fuels and chemicals is highly dependent on the structure and chemical composition of the biomass. So, rapid estimation of the chemical composition of biomass will be highly beneficial for biorefineries to fine-tune the feedstock and other processing parameters like enzymatic loading for a better yield of 2G ethanol. The conventional wet chemistry method for composition analysis is time-consuming, costly and laborious. This study aims to develop rapid, non-laborious, low-cost, and industrially applicable chemometric models like principal component regression (PCR) and partial least squares regression (PLSR) based on FTIR spectroscopy to determine the chemical composition of one such typical lignocellulosic biomass, viz., rice straw (RS). However, the results suggest that PCR and PLSR models constructed using the unprocessed FTIR spectra show poor performance in prediction. So, the spectra were processed, and an exploratory spectral analysis helped identify a specific region from 750 to 1800 cm−1 (SR-1) that accounts for significant variation between the RS samples. Moreover, 58 critical peaks in SR-1 were identified using a novel peak identification method proposed in this study. Further, results suggest that PLSR models developed using SR-1 of the processed spectra and the peaks as excellent prediction models (\({\mathrm{R}}_{\mathrm{cv}}^{2}\) > 95% and RPD > 4) and successful prediction models (90% < \({\mathrm{R}}_{\mathrm{cv}}^{2}\)< 95% and 3 < RPD < 4). Hence this study demonstrates that fine-tuned PLSR models based on processed FTIR spectra can be used as a tool for high-throughput screening of RS samples in biorefineries to improve the yield.Graphical Abstract
... The asymmetric least squares (ALS) smoothing approach introduced by Eilers and Boelens is adopted for all experiments [37]. The ALS method uses asymmetric weighting of deviations to obtain a baseline estimator, with the advantage of retaining the signal peak information, as can be seen in Fig. 4. The ALS method has three variable spectral descriptors: lambda, p, and Niterations. ...
... Lambda is the second derivative constraint, p is the weighting of positive residuals, and Niterations is the maximum number of iterations. Fig. 3 depicts the pseudocode of the ALS algorithm when adapted to Python [37]. More iterations produce a better signal but take more time. ...
... Pseudocode of the ALS algorithm when adapted to Python[37]. ...
In the recycling industry, the use of deep spectral convolutional networks for the purpose of material classification and composition estimation is still limited, despite the great opportunities of these techniques. In this study, the use of Laser-Induced Breakdown Spectroscopy (LIBS), Machine Learning (ML), and Deep Learning (DL) for the three-way sorting of Aluminum (Al) is proposed. Two sample sets of Al scrap are used: one containing 733 pieces for pre-training and validation with a ground truth of X-Ray Fluorescence (XRF), and the second containing 210 pieces for testing for unknown compositions. The proposed method comprises a denoising system combined with a method that extracts 145 features from the raw LIBS spectra. Further, three ML algorithms are assessed to identify the best-performing one to classify unknown pieces of aluminum post-consumer scrap into three commercially interesting output classes. The classified pieces are weighed, melted, and analyzed using spark analysis. Finally, to optimize the best-performing ML system, three state-of-the-art denoising and three feature extraction networks are pre-trained for learning the baseline correction and the proposed feature extraction. Transfer Learning from the six pre-trained networks is applied to create and evaluate 24 end-to-end DL models to classify Al in real-time from >200 spectra simultaneously. The end-to-end DL scheme shows the advantages of learning and denoising the spectra, allowing the transfer of traditional spectral analysis knowledge and the proposed feature extraction into DL, where the network learns from the entire spectrum. The best results for ML and DL were obtained with Random Forest processing one spectrum in 150 ms and BPNN+GHOSTNET(Fine-tuning) processing 200 spectra in 9 ms, which achieved 0.80 Precision, 0.81 Recall, 0.80 F1-score, and 0.80 Precision, 0.79 Recall, 0.79 F1-score, respectively.
... c-e, The results of three preprocessing pipelines built within RamanSPy, demonstrating the need for standardisation. Note on preprocessing methods: fingerprint region is 700-1800 cm −1 ; ASLS -Asymmetric Least Squares [50]; asPLS -Adaptive Smoothness Penalized Least Squares [51]; AUC -area under the curve; cosmic rays removed with algorithm from [52]. 4 RamanSPy interfaces with AI/ML Python frameworks to create new methods for RS analysis. a, RamanSPy allows users to incorporate AI/ML models seamlessly into pipelines created within the platform. ...
... (3) denoising with a Savitzky-Golay filter polynomial order 3 and kernel size 7 [56]; (4) baseline correction with asymmetric least squares [50]; and (5) Global MinMax normalisation to the interval [0, 1]. ...
Raman spectroscopy is a non-destructive and label-free chemical analysis technique, which plays a key role in the analysis and discovery cycle of various branches of science. Nonetheless, progress in Raman spectroscopic analysis is still impeded by the lack of software, methodological and data standardisation, and the ensuing fragmentation and lack of reproducibility of analysis workflows thereof. To address these issues, we introduce RamanSPy, an open-source Python package for Raman spectroscopic research and analysis. RamanSPy provides a comprehensive library of ready-to-use tools for spectroscopic analysis, which streamlines day-to-day tasks, integrative analyses, as well as novel research and algorithmic development. RamanSPy is modular and open source, not tied to a particular technology or data format, and can be readily interfaced with the burgeoning ecosystem for data science, statistical analysis and machine learning in Python.
... c-e, The results of three preprocessing pipelines built within RamanSPy, demonstrating the need for standardisation. Note on preprocessing methods: fingerprint region is 700-1800 cm −1 ; ASLS -Asymmetric Least Squares [50]; asPLS -Adaptive Smoothness Penalized Least Squares [51]; AUC -area under the curve; cosmic rays removed with algorithm from [52]. 4 RamanSPy interfaces with AI/ML Python frameworks to create new methods for RS analysis. a, RamanSPy allows users to incorporate AI/ML models seamlessly into pipelines created within the platform. ...
... (3) denoising with a Savitzky-Golay filter polynomial order 3 and kernel size 7 [56]; (4) baseline correction with asymmetric least squares [50]; and (5) Global MinMax normalisation to the interval [0, 1]. ...
Raman spectroscopy is a non-destructive and label-free chemical analysis technique, which plays a key role in the analysis and discovery cycle of various branches of science. Nonetheless, progress in Raman spectroscopic analysis is still impeded by the lack of software, methodological and data standardisation, and the ensuing fragmentation and lack of reproducibility of analysis workflows thereof. To address these issues, we introduce RamanSPy, an open-source Python package for Raman spectroscopic research and analysis. RamanSPy provides a comprehensive library of ready-to-use tools for spectroscopic analysis, which streamlines day-to-day tasks, integrative analyses, as well as novel research and algorithmic development. RamanSPy is modular and open source, not tied to a particular technology or data format, and can be readily interfaced with the burgeoning ecosystem for data science, statistical analysis and machine learning in Python.
... Following alignment, spectra are intensity normalised based on the sum of the metabolite-rich spectral data points between 4 and 0.2 ppm to reduce the influence of B1 inhomogeneity. Since broad signal components, such as residual water tails, can heavily influence normalisation, the Asymmetric Least Squares (ALS) baseline correction method ( Eilers and Boelens, 2005 ) is applied prior to summation. Note, the baseline corrected spectra are only used to derive the normalisation scaling values, which are subsequently applied to the uncorrected spectra. ...
... quential application of discrete spectral processing steps can achieve this aim. Partial volume effects; frequency and phase inconsistencies; baseline; and linewidth variability are reduced using a variety of techniques ( Eilers and Boelens, 2005 ;Goryawala et al., 2018 ;Wilson, 2019 ), and combined into a novel and fully automated analysis pipeline: SLIP-MAT. The method is validated using MRSI data from 8 healthy participants, highlighting subtle spectral features associated with individual differences in neuro-metabolism in strong agreement with previous work ( Wu et al., 2022 ). ...
1H Magnetic Resonance Spectroscopy (MRS) is an important non-invasive tool for measuring brain metabolism, with numerous applications in the neuroscientific and clinical domains. In this work we present a new analysis pipeline (SLIPMAT), designed to extract high-quality, tissue-specific, spectral profiles from MR spectroscopic imaging data (MRSI). Spectral decomposition is combined with spatially dependant frequency and phase correction to yield high SNR white and grey matter spectra without partial-volume contamination. A subsequent series of spectral processing steps are applied to reduce unwanted spectral variation, such as baseline correction and linewidth matching, before direct spectral analysis with machine learning and traditional statistical methods. The method is validated using a 2D semi-LASER MRSI sequence, with a 5-minute duration, from data acquired in triplicate across 8 healthy participants. Reliable spectral profiles are confirmed with principal component analysis, revealing the importance of total-choline and scyllo-inositol levels in distinguishing between individuals - in good agreement with our previous work. Furthermore, since the method allows the simultaneous measurement of metabolites in grey and white matter, we show the strong discriminative value of these metabolites in both tissue types for the first time. In conclusion, we present a novel and time efficient MRSI acquisition and processing pipeline, capable of detecting reliable neuro-metabolic differences between healthy individuals, and suitable for the sensitive neurometabolic profiling of in-vivo brain tissue.
... A method going from Whittaker (Eilers, 2004;Eilers, 2005) will be used here. Promising approach uses formulations similar to those of ill-posed problems (Eilers, 2003;Eilers and Boelens, 2005): ...
... Moreover, it is quite possible that the uniqueness of the solution is not ascertained. Therefore, the following approach is applied (Eilers 2004;Eilers and Boelens, 2005). First, a fixed large value of λ is established, based on experiments, and then the weights are iteratively set to two values, for example, 0.9 for the quiet parts of spectra and 0.1 for the peak areas. ...
How to process airborne gamma spectrometry data using models of gamma field in spectra. The processing has only three stages: extracting photopeaks, correcting for radon, and solving the inverse problems.
... The resonance term is related to the B-exciton and can be extracted from the UV-vis measurement (Figure 2.7) after applying a baseline correction [86] and fitting the resonances with Gaussian functions (Figure 2.36). The B-excitonic resonance was fit with a Gaussian centered at 525 nm with a FWHM of 9.8 nm. ...
... The FFTs in Figure 3.5 (b) have a lot of oscillations in the spectra that can obscure the absorption resonances and therefore need to be removed. The 0 % RH FFT in Figure 3.5 (b) was fit with a baseline [86] and an FFT was calculated to find the cause of the oscillations in the time-domain and the result is shown in original crystal had a faulty AR coating. As a result, the ZnTe time-domain signal can be measured without any significant reflections, and the spurious oscillations in the corresponding FFT are removed. ...
Ultrafast spectroscopies on the femtosecond timescale are well suited to study electronic dynamics and coupling between quasiparticles in condensed matter systems. These interactions are of great fundamental importance and have widespread applications in engineering. Future developments in spintronics, valleytronics, and optoelectronics depend on an understanding of the couplings of electrons, phonons, magnons, and plasmons. Quasiparticle interactions take place on ultrafast timescales and can be investigated at visible and far infrared wavelengths. This work describes the development and application of time- and frequency-resolved spectroscopies to several important systems, with the purpose of directly detecting the specific quasiparticles involved in the dynamics and determining the timescales of their interactions. First, transient absorption spectroscopy and pump-degenerate four-wave mixing are utilized together to observe the electron-phonon and phonon-phonon scattering in transition metal dichalcogenides. These two spectroscopies jointly identify the specific phonons involved in these interactions. The application of pump-dlegenerate four-wave mixing to a solid state material is novel and requires an in-depth discussion of the interpretation of the measurement. Second, a time-resolved and time-domain terahertz spectroscopy instrument is developed. The characterization of the system is provided and two individual studies show its capabilities. Magnon excitations in single crystal NiO are first measured with terahertz absorption spectroscopy. The determination of these magnon frequencies demonstrates the high frequency resolution of the instrument. Following the magnon study, the charge carrier lifetimes in GaAs and ErAs:GaAs heterostructures are measured with optical pump-terahertz probe spectroscopy. The instrumentation and models of the ultrafast dynamics provide a framework for studying quasiparticle interactions in solid state systems with these techniques.
... Asymmetric least squares (ALS) allows complex baselines to be eliminated, wh preserving the shape of the spectra [15]. It is a technique often used in Raman spectrom try to eliminate the fluorescence background. ...
... Asymmetric least squares (ALS) allows complex baselines to be eliminated, while preserving the shape of the spectra [15]. It is a technique often used in Raman spectrometry to eliminate the fluorescence background. ...
Even though NIR spectroscopy is based on the Beer–Lambert law, which clearly relates the concentration of the absorbing elements with the absorbance, the measured spectra are subject to spurious signals, such as additive and multiplicative effects. The use of NIR spectra, therefore, requires a preprocessing step. This article reviews the main preprocessing methods in the light of aquaphotomics. Simple methods for visualizing the spectra are proposed in order to guide the user in the choice of the best preprocessing. The most common chemometrics preprocessing are presented and illustrated by three real datasets. Some preprocessing aims to produce a spectrum as close as possible to the absorbance that would have been measured under ideal conditions and is very useful for the establishment of an aquagram. Others, dedicated to the improvement of the resolution of the spectra, are very useful for the identification of the peaks. Finally, special attention is given to the problem of reducing multiplicative effects and to the potential pitfalls of some very popular methods in chemometrics. Alternatives proposed in recent papers are presented.
... A baseline correction was performed in each Raman spectrum by Asymmetric Least Squares (ALS) [21] and Piecewise Linear Fitting (PLF) [22] for decreasing the noise and fluorescence background without removing the original Raman information of each spectrum. ...
... The implementation of ALS minimizes the residual (S) of the difference between the elements of the measured spectrum (y i ) and the elements of the baseline (z i ). Such a relation is given by Equation 1 [21] : ...
Niobium (Nb) and tantalum (Ta) concentrated in pyrochlore and microlite mineral groups, respectively, have attracted worldwide attention due to their importance to aerospace and electronics industries. This manuscript addresses the use of Raman spectroscopy coupled with artificial neural networks (ANNs) for improving the identification and characterization of mineral species belonging to pyrochlore and microlite mineral groups. Spectral data were collected in the 100‐1400 cm‐1 range and two baseline corrections, namely Asymmetric Least Squares (ALS) and Piecewise Linear Fitting (PLF) were performed and compared. In most cases, ALS achieved better performance in the removal of background noise with no elimination of important features of the original spectrum. The ANNs were fed with balanced datasets and based on different topologies with logistics, hyperbolic tangent, and rectified linear unit activation functions in the hidden layers.
... All remaining raw fluorescence time traces were baseline corrected by asymmetric least squares smoothing, 74 absolute fluorescence changes transformed into DF/F, 26 ...
Glomeruli are the functional units of the vertebrate olfactory bulb (OB) connecting olfactory receptor neuron (ORN) axons and mitral/tufted cell (MTC) dendrites. In amphibians, these two circuit elements regularly branch and innervate multiple, spatially distinct glomeruli. Using functional multiphoton-microscopy and single-cell tracing, we investigate the impact of this wiring on glomerular module organization and odor representations on multiple levels of the Xenopus laevis OB network. The glomerular odor map to amino acid odorants is neither stereotypic between animals nor chemotopically organized. Among the morphologically heterogeneous group of uni- and multi-glomerular MTCs, MTCs can selectively innervate glomeruli formed by axonal branches of individual ORNs. We conclude that odor map heterogeneity is caused by the coexistence of different intermingled glomerular modules. This demonstrates that organization of the amphibian main olfactory system is not strictly based on uni-glomerular connectivity.
... It was hoped that they might transform the data in each class to be more consistent and exacerbate differences among other classes. For logistic regression (LR), baseline removal (BLR) was tested using Fully Automated Baseline Correction (FABC) (Cobas et al., 2006;Kajfosz and Kwiatek, 1987), Adaptive Iteratively Reweighted Penalized Least Squares (AirPLS) (Zhang et al., 2010), Asymmetric Least Squares (ALS) (Eilers and Boelens, 2005), or Morphologically Weighted Penalized Least Squares (MPLS) (Li et al., 2013). These were found to improve prediction accuracy only slightly, from ~55% for no BLR to ~60% for an LR model. ...
... The binned count rate data are given in Figure 5a, showing the relative intensities of background and source energies. Figure 5b displays the count rate data with an asymmetric least-squares baseline (p = 0.001, λ = 0.0001) subtraction applied [19,20]. The peaks associated with each respective source are labelled. ...
The full instrument response of a CsI(Tl)-PiN photodiode radioactivity detector, intended for deployment on a meteorological radiosonde, has been modelled by combining a physics-based model of the sensor with the detector circuit response, obtained via a LTSpice simulation. The model uses the incident energy of a gamma ray as an input, and produces the pulse expected from the detector. The detector response was verified by comparing the simulated energy calibration with laboratory radioactive sources. The Schmitt trigger part of the measurement circuit is found to control the observed minimum detectable energy of 223 keV. Additionally, the energy sensitivity of the PiN detector was found to be 0.529 $\pm$ 0.010 mV/keV in the 200-800 keV range. The simulation and laboratory calibrations were consistent to better than 20% over the operating range of the instrument, decreasing to 0.34% at 800 keV.
... Compared with spectra in Fig. 5 b), the SNR of spectra in Fig. 5 a) is lower due to low signal level caused by the opacity of the emulsion. Time gate width of 810 ps (TDC bins 11-40, resolution 27 ps) was used for spectra in Fig. 5. Post-processing steps for spectra in Fig. 5 are dark count subtraction and baseline subtraction (DCR measured by the third measurement of the triple measurement mode, baseline estimated by the asymmetric least squares smoothing [17]). ...
The potential of in-line Raman spectrometers for process monitoring applications has been shown for many industrial processes, but in most cases, only one measurement point has been monitored by one spectrometer. In this letter, we describe and demonstrate a novel, time-resolved method for measuring Raman spectra and fluorescence lifetimes from multiple points using a single excitation source and a single spectrometer. This technique is based on a combination of a time-resolved CMOS SPAD (single-photon avalanche diode) line sensor and a fitting optical light guiding system. The line sensor is designed to make multiple individual measurements at intervals of tens of nanoseconds and the optical light guiding system, in turn, produces matching temporal differences for optical signals from different measurement points. Thereby, signals from different points are distinguished in the time domain. A combined Raman and fluorescence lifetime monitoring of two measurement points was demonstrated with an oil-ethanol emulsion sample.
... Each twenty successive time points were averaged to reduce noise. Furthermore, the slowly varying signal for each cell signal was determined by the method of Eilers and Boelens [36] and subtracted from the original signal. This resulted in a flat baseline. ...
The efficient and biocompatible transfer of nucleic acids into mammalian cells for research applications or medical purposes is a long-standing, challenging task. Viral transduction is the most efficient transfer system, but often entails high safety levels for research and potential health impairments for patients in medical applications. Lipo- or polyplexes are commonly used transfer systems but result in comparably low transfer efficiencies. Moreover, inflammatory responses caused by cytotoxic side effects were reported for these transfer methods. Often accountable for these effects are various recognition mechanisms for transferred nucleic acids. Using commercially available fusogenic liposomes (Fuse-It-mRNA), we established highly efficient and fully biocompatible transfer of RNA molecules for in vitro as well as in vivo applications. We demonstrated bypassing of endosomal uptake routes and, therefore, of pattern recognition receptors that recognize nucleic acids with high efficiency. This may underlie the observed almost complete abolishment of inflammatory cytokine responses. RNA transfer experiments into zebrafish embryos and adult animals fully confirmed the functional mechanism and the wide range of applications from single cells to organisms.
... Cosmic spikes (narrow intensive peaks in Raman spectra caused by high-energy photons hitting the highly sensitive CCD detector) were eliminated using an advanced median filter [34,35]. Baseline was corrected using an automated asymmetric least-squares (AsLS) algorithm [36]. The spectral noise was reduced by applying a Savitzky-Golay filter [37]. ...
Optical spectroscopic analysis of the chemical composition of milk in its natural state is complicated by a complex colloidal structure, represented by differently sized fat and protein particles. The classical techniques of molecular spectroscopy in the visible, near-, and mid-infrared ranges carry only bulk chemical information about a sample, which usually undergoes a destructive preparation stage. The combination of Raman spectroscopy with confocal microscopy provides a unique opportunity to obtain a vibrational spectrum at any single point of the sample volume. In this study, scanning confocal Raman microscopy was applied for the first time to investigate the chemical microstructure of milk using samples of various compositions. The obtained hyperspectral images of selected planes in milk samples are represented by three-dimensional data arrays. Chemometric data analysis, in particular the method of multivariate curve resolution, has been used to extract the chemical information from complex partially overlaid spectral responses. The results obtained show the spatial distribution of the main chemical components, i.e., fat, protein, and lactose, in the milk samples under study using intuitive graphical maps. The proposed experimental and data analysis method can be used in an advanced chemical analysis of natural milk and products on its basis.
... Video files were converted to raw mean grey values using ImageJ. Baseline fluorescence F 0 was calculated in MATLAB using the Asymmetric least square mean smoothing method (Eilers and Boelens, 2005). This method allowed us to find a variable baseline F 0 without any prior knowledge of the peak regions of the signal. ...
Introduction: Glioblastoma (GBM) invasiveness and ability to infiltrate deep into the brain tissue is a major reason for the poor patient prognosis for this type of brain cancer. Behavior of glioblastoma cells, including their motility, and expression of invasion-promoting genes such as matrix metalloprotease-2 (MMP2), are strongly influenced by normal cells found in the brain parenchyma. Cells such as neurons may also be influenced by the tumor, as many glioblastoma patients develop epilepsy. In vitro models of glioblastoma invasiveness are used to supplement animal models in a search for better treatments, and need to combine capability for high-throughput experiments with capturing bidirectional interactions between GBM and brain cells. Methods: In this work, two 3D in vitro models of GBM-cortical interactions were investigated. A matrix-free model was created by co-culturing GBM and cortical spheroids, and a matrix-based model was created by embedding cortical cells and a GBM spheroid in Matrigel. Results: Rapid GBM invasion occurred in the matrix-based model, and was enhanced by the presence of cortical cells. Little invasion occurred in the matrix-free model. In both types of models, presence of GBM cells resulted in a significant increase in paroxysmal neuronal activity. Discussion: Matrix-based model may be better suited for studying GBM invasion in an environment that includes cortical cells, while matrix-free model may be useful in investigation of tumor-associated epilepsy.
... Baseline variation is an important issue in many signal processing applications and can be addressed using baseline estimation or correction methods. NeoNNS benefits from an asymmetric least-squares smoothing (ALSS) correction algorithm (40) to automatically correct the nipple pressure signal baseline. The ALSS algorithm effectively pulls all the lower points of every nipple pressure waveform back to the zero baseline while maintaining the structure of the suck compression waveform shape. ...
Background: Non-nutritive suck (NNS) is used to promote ororhythmic patterning and assess oral feeding readiness in preterm infants in the neonatal intensive care unit (NICU). While time domain measures of NNS are available in real time at cribside, our understanding of suck pattern generation in the frequency domain is limited. The aim of this study is to model the development of NNS in the frequency domain using Fourier and machine learning (ML) techniques in extremely preterm infants (EPIs). Methods: A total of 117 EPIs were randomized to a pulsed or sham orocutaneous intervention during tube feedings 3 times/day for 4 weeks, beginning at 30 weeks post-menstrual age (PMA). Infants were assessed 3 times/week for NNS dynamics until they attained 100% oral feeding or NICU discharge. Digitized NNS signals were processed in the frequency domain using two transforms, including the Welch power spectral density (PSD) method, and the Yule-Walker PSD method. Data analysis proceeded in two stages. Stage 1: ML longitudinal cluster analysis was conducted to identify groups (classes) of infants, each showing a unique pattern of change in Welch and Yule-Walker calculations during the interventions. Stage 2: linear mixed modeling (LMM) was performed for the Welch and Yule-Walker dependent variables to examine the effects of gestationally-aged (GA), PMA, sex (male, female), patient type [respiratory distress syndrome (RDS), bronchopulmonary dysplasia (BPD)], treatment (NTrainer, Sham), intervention phase [1, 2, 3], cluster class, and phase-by-class interaction. Results: ML of Welch PSD method and Yule-Walker PSD method measures revealed three membership classes of NNS growth patterns. The dependent measures peak_Hz, PSD amplitude, and area under the curve (AUC) are highly dependent on PMA, but show little relation to respiratory status (RDS, BPD) or somatosensory intervention. Thus, neural regulation of NNS in the frequency domain is significantly different for each identified cluster (classes A, B, C) during this developmental period. Conclusions: Efforts to increase our knowledge of the evolution of the suck central pattern generator (sCPG) in preterm infants, including NNS rhythmogenesis will help us better understand the observed phenotypes of NNS production in both the frequency and time domains. Knowledge of those features of the Pediatric Medicine, 2023
... Others have proposed the use of discrete wavelet analysis to overcome challenges with noisy data [58]. In this work an alternative approach is taken, in which the AC and DC characteristics of the PPG waveforms are characterized using asymmetric least squares (ALS) fitting along the top and bottom of the signals, seen in the inset of Fig. 2A [59,60]. In this way, a pseudo-continuous measure is achieved. ...
Pulse oximetry is the most widespread method of monitoring heart rate and blood oxygen levels in both clinical and non-clinical settings. Current devices are ridged, bulky, and suitable only for short-term use. While the next generation of devices are moving toward wearable technologies, this focus on being unobtrusive is insufficient. To enable continuous high quality long-term monitoring, implanted devices are required. These will eliminate interference from light and sensitivity to skin pigmentation, allow enhanced performance during movement, and have no concerns around percussive damage. Here, an inexpensive, ultra-flexible pulse oximetry probe is demonstrated. The hybrid devices are fabricated on 5 µm Parylene C using laser ablation to define the circuit, and integrate small, rigid optoelectronic components. These are demonstrated in vivo on anesthetized pigs. Both transmission mode, and the novel reflection mode, are shown to be effective geometries for this. The heart rate measured by these devices shows < 2% variance from concurrent peripheral pulse oximetry measurements, along with an average variance of around 2.5% attributed predominantly to differences between central and peripheral oxygen saturations. In addition, it is shown that when implemented directly on the femoral artery, these devices record a more acute response to the variation in oxygen intake compared to the peripheral measurements. Finally, the same devices are shown to have the potential for use in monitoring venous oxygen content. This could open up the possibility of continuous monitoring of arteriovenous oxygen difference. These devices are straightforward to produce, biocompatible, and can be easily implanted during cardiovascular surgery, offering a route toward long-term implantation for continuous patient monitoring.
... Upon reactivation of the BO algorithm, the raw data exported by the ICS was further processed so that a consistent evaluation of the objective function was possible. This comprised a baseline correction using the asymmetric least squares method [29], with an asymmetry factor of 0.002 and a smoothing factor of 10 5 , which was found to give stable results. We note that this number might have to be adjusted depending on the problem at hand (sample, detector, etc.). ...
Contemporary complex samples require sophisticated methods for full analysis. This work describes the development of a Bayesian optimization algorithm for automated and unsupervised development of gradient programs. The algorithm was tailored to LC using a Gaussian process model with a novel covariance kernel. To facilitate unsupervised learning, the algorithm was designed to interface directly with the chromatographic system. Single-objective and multi-objective Bayesian optimization strategies were investigated for the separation of two complex (n>18, and n>80) dye mixtures. Both approaches found satisfactory optima in under 35 measurements. The multi-objective strategy was found to be powerful and flexible in terms of exploring the Pareto front. The performance difference between the single-objective and multi-objective strategy was further investigated using a retention modeling example. One additional advantage of the multi-objective approach was that it allows for a trade-off to be made between multiple objectives without prior knowledge. In general, the Bayesian optimization strategy was found to be particularly suitable, but not limited to, cases where retention modelling is not possible, although its scalability might be limited in terms of the number of parameters that can be simultaneously optimized.
... 2020b (Mathworks, Natick, Massachusetts, United States). Prior to analysis, spectra were subjected to (1) image size reduction, specifically shaped for each image, to reduce the size of the dataset; (2) removal of the part of the spectrum not useful for the analysis, consisting in the wavenumbers below 300 cm −1 and above 3,720 cm −1 approximately; (3) cosmic ray removal by use of median filtering (Matlab built-in function medfilt1 using default settings); (4) Alternating Least Squares (ALS) baseline correction according to Eilers and Boelens (2005), which has been shown to cope well with fluorescence contribution (De Juan et al., 2014), with parameters λ = 10 5 and p = 0.0005. Due to the heterogeneous distribution of wood polymers in the wood cell walls, the data were clustered using k-means cluster analysis (as implemented in Matlab), which successfully separated lignin rich parts of the cell wall, i.e., the cell corner and middle lamella (CCML), the cellulose rich secondary cell wall (S2), and the empty lumina of tracheids and ray cells (LUMEN). ...
Lopes, Dercilio Junior Verly, Gustavo Fardin Monti, Greg W. Burgreen, Jordao Cabral Moulin, Gabrielly dos Santos Bobadilha, Edward D. Entsminger, and Ramon Ferreira Oliveira. 2022. Creating high-resolution microscopic cross-section images of hardwood species using generative adversarial networks. Pages 23–32 in Thygesen, L.G., Egea, G., Bucksch, A., eds. Innovative use of imaging techniques within plant science. Lausanne: Frontiers Media SA. Fronters in Plant Science. Frontiers e-book, comprising all the articles featured in Research Topic. ISSN 1664-8714. ISBN 978-2-83250-952-4. DOI 10.3389/978-2-83250-952-4. https://www.frontiersin.org/research-topics/34975/innovative-use-of-imaging-techniques-within-plant-science#articles Published December 22, 2022.
... For spectra, experimental data are processed in Interactive Data Language (IDL), and the continuum is fit with ACOFI (automated continuum fitter), which uses an asymmetric least squares fitting method that has been, in particular, adapted to the problem of continuum fitting in absorption spectra. 38 Subsequently, the experimental transmission data are compared to simulated transmission spectra we produced using the SPECTRUM code, 9,12,33 which is a ray-trace code modeling radiation transport with self-emission in the optically thin approximation. We compare the processed spectra in Sec. ...
Predicting and modeling the behavior of experiments with radiation waves propagating through low-density foams require a detailed quantification of the numerous uncertainties present. In regimes where a prominent radiative shock is produced, key dynamical features include the shock position, temperature, and curvature and the spatial distribution and temperature of the corresponding supersonic radiation wave. The COAX experimental platform is designed to constrain numerical models of such a radiative shock propagating through a low-density foam by employing radiography for spatial and shock information, Dante for characterizing the x-ray flux from the indirectly driven target, and a novel spectral diagnostic designed to probe the temperature profile of the wave. In this work, we model COAX with parameterized 2D simulations and a Hohlraum-laser modeling package to study uncertainties in diagnosing the experiment. The inferred temperature profile of the COAX radiation transport experiments has been shown to differ from simulations more than expected from drive uncertainties that have been constrained by simultaneous soft x-ray flux and radiography measurements.
... Spectra were trimmed to the 1000 data points in the 800−1800 cm −1 fingerprint region of most interest, then preprocessed with a background correction using the Asymmetric Least Squares Smoothing (ALSS) baselining algorithm. 21 This was followed by average normalization. ...
We report a novel method with higher than 90% accuracy in diagnosing buccal mucosa cancer. We use Fourier transform infrared spectroscopic analysis of human serum by suppressing confounding high molecular weight signals, thus relatively enhancing the biomarkers' signals. A narrower range molecular weight window of the serum was also investigated that yielded even higher accuracy on diagnosis. The most accurate results were produced in the serum's 10-30 kDa molecular weight region to distinguish between the two hardest to discern classes, i.e., premalignant and cancer patients. This work promises an avenue for earlier diagnosis with high accuracy as well as greater insight into the molecular origins of these signals by identifying a key molecular weight region to focus on.
... En el cas de les dades cromatogràfiques, es va utilitzar l'algoritme de mínim quadrats asimètrics (de l'anglès Asymmetric Least-Squares, AsLS) (Eilers & Boelens, 2005) per a corregir els possibles canvis en la línia de base dels cromatogrames. Aquest mètode és especialment adequat per a corregir les línies de base d'aquells conjunts de dades en què el senyal significatiu (pics cromatogràfics) és molt més estret que la contribució ampla de la línia de base, com succeeix normalment en les dades cromatogràfiques (Eilers, 2004). ...
In recent years, the increase in the production of synthetic organic compounds and their constant release into the environment, especially in reference to pharmaceuticals, compounds on which this thesis has focused, has meant that these substances of origin generally anthropogenic have been accumulating in biological ecosystems in an important, extensive, and persistent way, especially in the aquatic environment. For this reason, it is necessary to use environmental purification processes to eliminate substances that are harmful to health and the environment. However, the purification technologies used today do not yet allow these substances to be completely eliminated in a quantitative way. In addition, the potentially harmful effects of these substances on the environment are not only due to the original compounds, but in many cases also to their degradation products and metabolites. Consequently, it is important to know the fate and transformation that organic pollutants undergo once they enter the ecosystem and reach the aquatic environment. We must therefore keep in mind that water is an essential resource for life that requires the best possible sanitary quality for its consumption. Dissolved organic matter (DOM), present in both surface and groundwater, and which has a significant effect on the biochemical processes of aquatic systems, often acts as an indicator of the level of quality and health of water. Therefore, when optimizing the purification processes carried out, it is also very important to know and act on the concentration and characteristics of the DOM present in it. For this reason, the following studies have been performed in this Thesis: On the one hand, to better understand the process of photodegradation of drugs, the degradation process of two of them has been studied. In a first study, different samples of a cytostatic drug (tamoxifen) were subjected to a controlled source of UV radiation in the laboratory, which simulates solar radiation (suntest), using two different irradiation powers. At the same time, its photodegradation is monitored using UV-Visible spectrophotometric measurements and multidimensional fluorescence (excitation-emission matrices, EEM). Then, different sample aliquots along the degradation are analyzed by LC-DAD, LC-MS and / or LC-FLD. These analyzes, which generate a very high amount of data, are merged and treated simultaneously using advanced chemometric procedures (especially with multivariate curve resolution alternating least squares method (MCR-ALS)), which describe the process and the degradation reaction, solve the different transformation products (TPs), identify them and evaluate the kinetics associated with the process studied. As a result of this study, 3 TPs and 1 isomer of tamoxifen were characterized. A second study with the same aim was carried out on the antibiotic sulfamethoxazole, present in rivers and effluents from all over Europe in remarkable concentrations. Additionally, a study of the acid-base properties of this substance was conducted to see how pH can affect the speciation of this compound. Similarly, to know whether the photodegradation of the drug is influenced by this parameter, the degradation process was studied using sample solutions at different pH. As a result of the chemometric analysis of the data from this second study, a conformational isomer, an unknown photoproduct, and 4 TPs were characterized. On the other hand, 3 different forms of the antibiotic were detected in the corresponding acid-base titration. On the other hand, to know the quality of the surface water of the rivers, different samples of surface water from the Llobregat river basin have been analyzed and characterized by multidimensional fluorescence (EEM). In this study, the monitoring of water quality from the combination of the fluorometric measurements of EEM and its flexible chemometric modeling, also through the MCR-ALS, has made it possible to obtain information on the nature of the fractions of the DOM (humic, fulvic, protein origin, etc.) present in the Llobregat river and its geographical distribution along the basin in different environmental sampling campaigns. Therefore, given that there is an urgent need today to develop new multivariate data processing tools, such as those obtained in this Thesis in the different cases of environmental interest studied, the presented report aims to contribute to solve this bottleneck from the use of chemometrics, and especially the MCR-ALS method, which can play a very important role in solving many of the problems associated with data analysis in different areas of chemical and environmental application.
... Spectral data were collected using either 0.25 or 2.5 mW laser power depending on the sensitivity of the sample to phase transformation. All Raman data were baseline-corrected with an asymmetric least squares algorithm (Eilers and Boelens, 2005) and min-max normalized before spectral fitting for feature identification. They were smoothed with the Savitzky-Golay algorithm for visual representation, but no features were assigned based on smoothed data. ...
Mineral transformations by two hyperthermophilic Fe(III)-reducing crenarchaea, Pyrodictium delaneyi and Pyrobaculum islandicum, were examined using synthetic nanophase ferrihydrite, lepidocrocite, and akaganeite separately as terminal electron acceptors and compared with abiotic mineral transformations under similar conditions. Spectral analyses using visible-near-infrared, Fourier-transform infrared attenuated total reflectance (FTIR-ATR), Raman, and Mössbauer spectroscopies were complementary and revealed formation of various biomineral assemblages distinguishable from abiotic phases. The most extensive biogenic mineral transformation occurred with ferrihydrite, which formed primarily magnetite with spectral features similar to biomagnetite relative to a synthetic magnetite standard. The FTIR-ATR spectra of ferrihydrite bioreduced by P. delaneyi also showed possible cell-associated organics such as exopolysaccharides. Such combined detections of biomineral assemblages and organics might serve as biomarkers for hyperthermophilic Fe(III) reduction. With lepidocrocite, P. delaneyi produced primarily a ferrous carbonate phase reminiscent of siderite, and with akaganeite, magnetite and a ferrous phosphate phase similar to vivianite were formed. P. islandicum showed minor biogenic production of a ferrous phosphate similar to vivianite when grown on lepidocrocite, and a mixed valent phosphate or sulfate mineral when grown on akaganeite. These results expand the range of biogenic mineral transformations at high temperatures and identify spacecraft-relevant spectroscopies suitable for discriminating mineral biogenicity.
... 2020b (Mathworks, Natick, Massachusetts, United States). Prior to analysis, spectra were subjected to (1) image size reduction, specifically shaped for each image, to reduce the size of the dataset; (2) removal of the part of the spectrum not useful for the analysis, consisting in the wavenumbers below 300 cm −1 and above 3,720 cm −1 approximately; (3) cosmic ray removal by use of median filtering (Matlab built-in function medfilt1 using default settings); (4) Alternating Least Squares (ALS) baseline correction according to Eilers and Boelens (2005), which has been shown to cope well with fluorescence contribution (De Juan et al., 2014), with parameters λ = 10 5 and p = 0.0005. Due to the heterogeneous distribution of wood polymers in the wood cell walls, the data were clustered using k-means cluster analysis (as implemented in Matlab), which successfully separated lignin rich parts of the cell wall, i.e., the cell corner and middle lamella (CCML), the cellulose rich secondary cell wall (S2), and the empty lumina of tracheids and ray cells (LUMEN). ...
Water is a key element for wood performance, as water molecules interact with the wood structure and affect important material characteristics such as mechanical properties and durability. Understanding wood-water interactions is consequently essential for all applications of wood, including the design of wood materials with improved durability by chemical modification. In this work, we used Raman micro-spectroscopy in combination with a specially designed moisture chamber to map molecular groups in wood cell walls under controlled moisture conditions in the hygroscopic range. We analyzed both untreated and chemically modified (acetylated to achieve two different spatial distributions of acetyl groups within the cell wall) Norway spruce wood. By moisture conditioning the specimens successively to 5, 50, and 95% relative humidity using deuterium oxide (D 2 O), we localized the moisture in the cell walls as well as distinguished between hydroxyl groups accessible and inaccessible to water. The combination of Raman micro-spectroscopy with a moisturizing system with deuterium oxide allowed unprecedented mapping of wood-water interactions. The results confirm lower moisture uptake in acetylated samples, and furthermore showed that the location of moisture within the cell wall of acetylated wood is linked to the regions where acetylation is less pronounced. The study demonstrates the local effect that targeted acetylation has on moisture uptake in wood cell walls, and introduces a novel experimental set-up for simultaneously exploring sub-micron level wood chemistry and moisture in wood under hygroscopic conditions.
... All data analysis was performed in locally written algorithms in MATLAB environment. To remove tissue autofluorescence, spectra were subject to baseline subtraction using a Whittaker filter with an asymmetric least squares algorithm [63][64][65] . This was followed by cosmic ray removal [66] , smoothing using a 2nd order, 5 point window Savitzky-Golay filter to reduce random spectral variations related to noise, and Lipid and protein C -H, CH 2 and CH 3 stretch vibrations [ 17 , 73 ] relative intensity correction using NIST SRM 2241 (National Institute of Standards and Technology) [67] . ...
The development of novel biomaterials for regenerative therapy relies on the ability to assess tissue development, quality, and similarity with native tissue types in in vivo experiments. Non-invasive imaging modalities such as X-ray computed tomography offer high spatial resolution but limited biochemical information while histology and biochemical assays are destructive. Raman spectroscopy is a non-invasive, label-free and non-destructive technique widely applied for biochemical characterization. Here we demonstrate the use of fibre-optic Raman spectroscopy for in vivo quantitative monitoring of tissue development in subcutaneous calcium phosphate scaffolds in mice over 16 weeks. Raman spectroscopy was able to quantify the time dependency of different tissue components related to the presence, absence, and quantity of mesenchymal stem cells. Scaffolds seeded with stem cells produced 3-5 times higher amount of collagen-rich extracellular matrix after 16 weeks implantation compared to scaffolds without. These however, showed a 2.5 times higher amount of lipid-rich tissue compared to implants with stem cells. Ex vivo micro-computed tomography and histology showed stem cell mediated collagen and bone development. Histological measures of collagen correlated well with Raman derived quantifications (correlation coefficient in vivo 0.74, ex vivo 0.93). In the absence of stem cells, the scaffolds were largely occupied by adipocytes. The technique developed here could potentially be adapted for a range of small animal experiments for assessing tissue engineering strategies at the biochemical level.
We report mechanochemically-induced deuteration of Pd-activated aromatic C(sp2)–H bonds at ambient temperature under solvent-free conditions. Deuterium was sourced from cysteine-d4 to obtain mono- or dideuterated products from various aromatic palladacycles....
Breast cancer diagnosis is crucial for timely treatment and improved outcomes. This paper proposes a novel approach for rapid breast cancer diagnosis using optical fiber probe-based attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy from 750 to 4000 cm-1 . The technique enables direct analysis of tissue samples, eliminating the need for microtome sectioning and staining, thus saving time and resources. By capturing molecular fingerprint information, various machine learning models were used to analyze the spectroscopic data to classify cancerous and non-cancerous tissues accurately. Comparing deparaffinized and paraffinized samples reveals the impact of sample preparation and experimental methods. The study demonstrates a strong correlation between the cancerous nature of a sample and its ATR-FTIR spectrum, suggesting its potential for breast cancer diagnosis (sensitivity of 74.2% and specificity of 78.3%). The proposed approach holds promise for integration into clinical operations, providing a rapid method for preliminary breast cancer diagnosis. This article is protected by copyright. All rights reserved.
The precipitation of struvite, a magnesium ammonium phosphate hexahydrate (MgNH4PO4⋅6H2O) mineral, from wastewater is a promising method for recovering phosphorous. While this process is commonly used in engineered environments, our understanding of the underlying mechanisms responsible for the formation of struvite crystals remains limited. Specifically, indirect evidence suggests the involvement of an amorphous precursor and the occurrence of multi-step processes in struvite formation, which would indicate non-classical paths of nucleation and crystallization. In this study, we use synchrotron-based in situ X-ray scattering complemented by cryogenic transmission electron microscopy to obtain new insights from the earliest stages of struvite formation. The holistic scattering data captured the structure of an entire assembly in a time-resolved manner. The structural features comprise the aqueous medium, the growing struvite crystals, and any potential heterogeneities or complex entities. By analysing the scattering data, we found that the onset of crystallization causes a perturbation in the structure of the surrounding aqueous medium. This perturbation is characterized by the occurrence and evolution of Ornstein-Zernike fluctuations on a scale of about 1 nm, suggesting a non-classical nature of the system. We interpret this phenomenon as a liquid-liquid phase separation (LLPS), which gives rise to the formation of the amorphous precursor phase preceding actual crystal growth of struvite. Our microscopy results confirm that the formation of Mg-struvite includes a short-lived amorphous phase, lasting >10 seconds.
Electronic cigarettes are a popular nicotine consumption product that have risen in popularity as an alternative to cigarettes. However, their recent meteoric rise in market size and various controversies have resulted in the analyses of e-liquid ingredients to be focused on powerful laboratory-based slow methods such as chromatography and mass spectrometry. Here we present a complementary technology based on Raman spectroscopy combined with chemometrics as a fast, inexpensive, and highly portable screening tool to detect and quantify the propylene glycol : glycerol (PG : VG) ratio and nicotine content of e-cigarette liquids. Through this, the PG : VG ratio of 20 out of 23 commercial samples was quantified to within 3% of their stated value, while nicotine was successfully quantified to within 1 mg g-1 for 16 out of 23 samples without the need for accurate knowledge of flavonoid composition. High linearity was also achieved when flavours were kept constant. Finally, the limitations of Raman spectroscopy are discussed, and potential solutions are suggested.
To safeguard the quality of river water, a comprehensive approach is required within the European Water Framework Directive. It is vital to conduct non-target screening of the complete chemical fingerprint of the aquatic ecosystem, as this will help to identify chemicals of emerging concern and uncover their unusual dynamic patterns in river water. Achieving this goal calls for an advanced combination of two measurement paradigms: tracing the potential pollution path through the river network and detecting the numerous compounds that constitute the chemical composition, both known and unknown. To address this challenge, we propose an integrated approach that combines the preprocessing of ongoing Gas Chromatography Mass Spectrometry (GC-MS) measurements at nine sites along the Rhine using PARAllel FActor Analysis2 (PARAFAC2) for non-target screening, with spatiotemporal modelling of these sites within the river network using a statistical path modelling algorithm called Process Partial Least Squares (Process PLS). With an average explained variance of 97.0%, PARAFAC2 extracted mass spectra, elution, and concentration profiles of known and unknown chemicals. On average, 76.8% of the chemical variability captured by the PARAFAC2 concentration profiles was extracted by Process PLS. The integrated approach enabled us to track chemicals through the Rhine catchment, and tentatively identify known and as-yet unknown potential pollutants, including methyl tert-butyl ether and 1,3-cyclopentadiene, based on non-target screening and spatiotemporal behaviour.
Identifying fungal clones propagated during outbreaks in hospital settings is a problem that increasingly confronts biologists. Current tools based on DNA sequencing or microsatellite analysis require specific manipulations that are difficult to implement in the context of routine diagnosis. Using deep learning to classify the mass spectra obtained during the routine identification of fungi by MALDI-TOF mass spectrometry could be of interest to differentiate isolates belonging to epidemic clones from others. As part of the management of a nosocomial outbreak due to Candida parapsilosis in two Parisian hospitals, we studied the impact of the preparation of the spectra on the performance of a deep neural network. Our purpose was to differentiate 39 otherwise fluconazole-resistant isolates belonging to a clonal subset from 56 other isolates, most of which were fluconazole-susceptible, collected during the same period and not belonging to the clonal subset. Our study carried out on spectra obtained on four different machines from isolates cultured for 24 or 48 h on three different culture media showed that each of these parameters had a significant impact on the performance of the classifier. In particular, using different culture times between learning and testing steps could lead to a collapse in the accuracy of the predictions. On the other hand, including spectra obtained after 24 and 48 h of growth during the learning step restored the good results. Finally, we showed that the deleterious effect of the device variability used for learning and testing could be largely improved by including a spectra alignment step during preprocessing before submitting them to the neural network. Taken together, these experiments show the great potential of deep learning models to identify spectra of specific clones, providing that crucial parameters are controlled during both culture and preparation steps before submitting spectra to a classifier.
Signal transduction and cell function are governed by the spatiotemporal organization of membrane-associated molecules. Despite significant advances in visualizing molecular distributions by 3D light microscopy, cell biologists still have limited quantitative understanding of the processes implicated in the regulation of molecular signals at the whole cell scale. In particular, complex and transient cell surface morphologies challenge the complete sampling of cell geometry, membrane-associated molecular concentration and activity and the computing of meaningful parameters such as the cofluctuation between morphology and signals. Here, we introduce u-Unwrap3D, a framework to remap arbitrarily complex 3D cell surfaces and membrane-associated signals into equivalent lower dimensional representations. The mappings are bidirectional, allowing the application of image processing operations in the data representation best suited for the task and to subsequently present the results in any of the other representations, including the original 3D cell surface. Leveraging this surface-guided computing paradigm, we track segmented surface motifs in 2D to quantify the recruitment of Septin polymers by blebbing events; we quantify actin enrichment in peripheral ruffles; and we measure the speed of ruffle movement along topographically complex cell surfaces. Thus, u-Unwrap3D provides access to spatiotemporal analyses of cell biological parameters on unconstrained 3D surface geometries and signals.
Each element on the periodic table has its own unique peak spectrum in LIBS analysis. Therefore, the chemical composition of a sample can be determined by identifying the wavelength of each peak from the analyzed sample. This has been a common sample identification method. However, due to various measurement conditions and environments, this peak search with database reference is a very tedious task. In this study, we propose a new method of analyzing LIBS spectra using only standard samples and no database reference or peak wavelength search, based on a machine learning method.
This research work focuses on the microphysical characterization of ice present on the surface of different planetary bodies. Studying the microphysical state of the ice consists in characterizing the chemical and structural properties such as the volume proportion of the components, the size of their grains, the porosity or the surface roughness. These properties allow us to understand the formation mode and the processes governing the temporal evolution of these surfaces. At the end of this thesis, two objects were the target of these investigations: Mars and Europa, a satellite of Jupiter. To do so, we use the physical principles describing the interactions between light and matter and in particular the theoretical framework of radiative transfer. The radiation transfer equations allow to model the phenomena of absorption, reflection and scattering of light during its interaction with the surface materials.This work is organized in three distinct parts. The first part is devoted to the description of radiation transfer, the concepts and important physical quantities are defined to lead to some examples of models commonly used by the scientific community. The model used in this work, the so-called photometric Hapke model, is widely described. The advantages of this approach lie in the fact that this model is analytical and easily invertible. The inversion of a physical model consists in using it to find the physical parameters allowing to reproduce as accurately as possible a data, in our case an observation of the surface. The concepts and methods of inversions are also presented in this section. A comparison between different inversion methods is proposed to select the most suitable methods for the problems encountered.The second part is dedicated to the characterization of the surface of Europa, one of the ice satellites of Jupiter. The scientific background gives an overview of the current knowledge about the properties of the surface and ends with the remaining questions: what is the chemical composition of the surface? How does it vary from one geological structure to another? What processes promote this composition? To answer these questions we use the data obtained by the NIMS spectro-imager during the Galileo mission. We have combined the Hapke model and a Bayesian inversion approach to test, for the first time, a very large number of different representations of the surface from 15 chemical compounds that have been proposed so far by previous studies. We show that there is a multitude of different surface representations that produce a similar fit to the data.The third part is devoted to the study of the ices of Mars via the use of the data of the recent ExoMars-TGO mission and in particular the NOMAD infrared spectrometer. The acquisition of the first data of the mission coincides with the beginning of this thesis, so an instrumental calibration work was necessary. A calibration method for the NOMAD nadir channel data has been proposed and published. A first analysis of the data allows to highlight the instrumental capacity to detect surface ice. These data will then allow to undertake a work of characterization and temporal follow-up of the microphysical properties of the ice on Mars.
One of the problems that most affect hospitals is infections by pathogenic microorganisms. Rapid identification and adequate, timely treatment can avoid fatal consequences and the development of antibiotic resistance, so it is crucial to use fast, reliable, and not too laborious techniques to obtain quick results. Raman spectroscopy has proven to be a powerful tool for molecular analysis, meeting these requirements better than traditional techniques. In this work, we have used Raman spectroscopy combined with machine learning algorithms to explore the automatic identification of eleven species of the genus Candida, the most common cause of fungal infections worldwide. The Raman spectra were obtained from more than 220 different measurements of dried drops from pure cultures of each Candida species using a Raman Confocal Microscope with a 532 nm laser excitation source. After developing a spectral preprocessing methodology, a study of the quality and variability of the measured spectra at the isolate and species level, and the spectral features contributing to inter-class variations, showed the potential to discriminate between those pathogenic yeasts. Several machine learning and deep learning algorithms were trained using hyperparameter optimization techniques to find the best possible classifier for this spectral data, in terms of accuracy and lowest possible overfitting. We found that a one-dimensional Convolutional Neural Network (1-D CNN) could achieve above 80% overall accuracy for the eleven classes spectral dataset, with good generalization capabilities.
One of the most promising innovation strategies for sorting and recycling post-consumer aluminium scrap is using quantitative Laser-Induced Breakdown Spectroscopy (LIBS) analysis. However, existing methods to estimate alloying element concentrations based on LIBS spectra, such as linear univariate regression and Machine Learning models, are still too limited in their performance to achieve the accuracy demanded by the industry. Therefore, this study presents novel Deep Learning approaches and compares their performance to those of traditional univariate regression and Machine Learning methods in terms of RMSE, MAE, and R2 value. For this evaluation, two sample sets of aluminium pieces are used: one containing 27 certified aluminium reference samples and the second containing 733 post-consumer scrap pieces for which the ground truth concentrations are determined by X-Ray Fluorescence (XRF).
Adopting multiple loss functions, one for each element has proven its significant value for the regression performance. It improves the results for all performance metrics in the Scrap Sample set, and the same is true for the Reference Sample set, except for the coefficient of determination of Fe, Mn and Mg. In addition, the proposed methodology considers the learning prioritisation problem to prevent that learning the concentration of the base element is prioritised over the alloying elements. Although the effect of excluding the base alloy aluminium from the learning is small and not always positive for the performance, demonstrating this effect is also considered valuable. Since the average RMSE on the prediction is just 0.02 wt.% for Al and Si, and not more than 0.01 wt.% for Fe, Cu, Mn, Mg, and Zn, the best performing Deep Learning model shows promise for the future of LIBS in metal sorting applications.
Raman spectroscopy was compared with near infrared (NIR) hyperspectral imaging for determination of fat composition (%EPA + DHA) in salmon fillets at short exposure times. Fillets were measured in movement for both methods. Salmon were acquired from several different farming locations in Norway with different feeding regimes, representing a realistic variation of salmon in the market. For Raman, we investigated three manual scanning strategies; i) line scan of loin, ii) line scan of belly and iii) sinusoidal scan of belly at exposure times of 2s and 4s. NIR images were acquired while the fillets moved on a conveyor belt at 40 cm/s, which corresponds to an acquisition time of 1s for a 40 cm long fillet. For NIR images, three different regions of interest (ROI) were investigated including the i) whole fillet, ii) belly segment, and iii) loin segment. For both Raman and NIR measurements, we investigated an untrimmed and trimmed version of the fillets, both relevant for industrial in-line evaluation. For the trimmed fillets, a fat rich deposition layer in the belly was removed. The %EPA + DHA models were validated by cross validation (N = 51) and using an independent test set (N = 20) which was acquired in a different season. Both Raman and NIR showed promising results and high performances in the cross validation, with R2CV = 0.96 for Raman at 2s exposure and R2CV = 0.97 for NIR. High performances were obtained also for the test set, but while Raman had low and stable biases for the test set, the biases were high and varied for the NIR measurements. Analysis of variance on the squared test set residuals showed that performance for Raman measurements were significantly higher than NIR at 1% significance level (p = 0.000013) when slope-and-bias errors were not corrected, but not significant when residuals were slope-and-bias corrected (p = 0.28). This indicated that NIR was more sensitive to matrix effects. For Raman, signal-to-noise ratio was the main limitation and there were indications that Raman was close to a critical sample exposure time at the 2s signal accumulation.
The ongoing worldwide emergence of epidemic azole-resistant Candida parapsilosis clones is a threat to human health. In order to monitor a clonal outbreak which involved 2 hospitals in Paris, we used a technology that most clinical laboratories are equipped with: Matrix-Assisted Laser Desorption-Ionization Time of Flight mass spectrometry (MALDI-TOF MS). This physico-chemical analysis technique was combined with machine learning algorithm to evaluate the possibility of monitoring a multi-center epidemic. A total of 96 isolates of Candida parapsilosis constituted of resistant clones, susceptible clones and other susceptible isolates were cultured on various culture conditions. Spectra were acquired on four different mass spectrometers from three hospitals.
We used convolutional neural networks (CNN) to classify the isolates and identified the main parameters that can influence the identification of epidemic clonal isolates. We have shown results up to 94% accuracy in a multi-center context, and between 81 and 91% accuracy when the tested site is not part of the training set, showing that it is possible to differentiate the spectra of a clone and a non-clone presenting a very strong similarity. This study shows that it is possible to monitor a certain type of clone, namely the R2 clones in our case study, using MALDI-TOF coupled to CNN. The experimental conditions that impacted the results should be considered for future work to improve performances for difficult classification tasks like clonal fungal identification.
Neural activity in the sensory cortex combines stimulus responses and
ongoing activity, but it remains unclear whether these reflect the same
underlying dynamics or separate processes. In the present study, we show
in mice that, during wakefulness, the neuronal assemblies evoked by sounds
in the auditory cortex and thalamus are specific to the stimulus and distinct
from the assemblies observed in ongoing activity. By contrast, under three
different anesthetics, evoked assemblies are indistinguishable from ongoing
assemblies in the cortex. However, they remain distinct in the thalamus. A
strong remapping of sensory responses accompanies this dynamic state
change produced by anesthesia. Together, these results show that the awake
cortex engages dedicated neuronal assemblies in response to sensory
inputs, which we suggest is a network correlate of sensory perception.
Three-dimensional (3D) printing is becoming an attractive technology for the design and development of personalized paediatric dosage forms with improved palatability. In this work micro-extrusion based printing was implemented for the fabrication of chewable paediatric ibuprofen (IBU) tablets by assessing a range of front runner polymers in taste masking. Due to the drug-polymer miscibility and the IBU plasticization effect, micro-extrusion was proved to be an ideal technology for processing the drug/polymer powder blends for the printing of paediatric dosage forms. The printed tablets presented high printing quality with reproducible layer thickness and a smooth surface. Due to the drug-polymer interactions induced during printing processing, IBU was found to form a glass solution confirmed by differential calorimetry (DSC) while H-bonding interactions were identified by confocal Raman mapping. IBU was also found to be uniformly distributed within the polymer matrices at molecular level. The tablet palatability was assessed by panellists and revealed excellent taste masking of the IBU’s bitter taste. Overall micro-extrusion demonstrated promising processing capabilities of powder blends for rapid printing and development of personalised dosage forms.
The exponential increase in data produced over the last two decades has revolutionized the way we collect, store, process, analyze, model, and interpret information to improve profitability. Manufacturing is no exception. However, Smart Manufacturing, the digital practice, organization, workforce, and infrastructure transformation for collection and deployment of data and models at scale and at all levels of manufacturing, is a complex, costly, and labor-intensive journey that is still seeing slow adoption. The Clean Energy Smart Manufacturing Innovation Institute (CESMII), a national Manufacturing USA public-private partnership sponsored by the Department of Energy, is addressing this scaled use of data and modeling in manufacturing. CESMII has focused on how to collect and use operating data for numerous applications that improve productivity, precision, and performance of manufacturing operations from factory floor to supply chain using process simulation, predictive analytics, monitoring and control, and real-time optimization. Because contextualized data are key, CESMII has developed the Smart Manufacturing Innovation Platform (SMIP) to lower the barriers to the data that are needed to accelerate data-based model building, improve data visualization, and more quickly gain insights. Reusable, standards-based ways of doing data collection, ingestion, and contextualization are particularly important for scaling access and use of data. The SMIP uses a standards-based definition and construct for reusable information models called an SM Profile. When an SM Profile is used in conjunction with the SMIP, the SMIP ensures the availability of contextualized, operational data for model building. The present work demonstrates Smart Manufacturing and the application of the SMIP for building several data-centered models for the operation and control of an experimental electrochemical reactor that reduces carbon dioxide (CO2) gas to valuable liquid and gas chemicals, such as alcohols, olefins, and syngas. We describe how the SMIP plays a central role in more effective model building and we demonstrate how the electochemical reactor can be controlled and optimized for the desired products. Use of the SMIP involves the transmission of real-time sensor measurements to a cloud resource so that the operating data are available to all model building experts. The data collection and transmission process is fully automated to greatly reduce the need for manual manipulation of the data. Data-driven machine learning models are used for advanced real-time state estimation, real-time optimization, and model-based feedback control for the reactor. The application models are implemented as a system to monitor the data flow and control the electrochemical reactor with a single visualization interface. SM Profile are used to demonstrate reusability of the information models for the the reactor and the instrumentation. The application packages, algorithms, and user interfaces developed are cast as Docker images in a library to facilitate reusability of the application models.
The detection and quantification of glucose concentrations in human blood or in the ocular fluid gain importance due to the increasing number of diabetes patients. A reliable determination of these low concentrations is hindered by the complex aqueous environments in which various biomolecules are present. In this study, we push the detection limit as well as the discriminative power of plasmonic nanoantenna-based sensors towards the physiological limit. We utilize plasmonic surface-enhanced infrared absorption spectroscopy (SEIRA) to study aqueous solutions of mixtures of up to five different physiologically relevant saccharides, namely the monosaccharides glucose, fructose, and galactose, as well as the disaccharides maltose and lactose. Resonantly tuned plasmonic nanoantennas in a reflection flow cell geometry allow us to enhance the specific vibrational fingerprints of the mono- and disaccharides. The obtained spectra are analyzed via principal component analysis (PCA) using a machine learning algorithm. The high performance of the sensor together with the strength of PCA allows us to detect concentrations of aqueous mono- and disaccharides solutions down to the physiological levels of 1 g/L. Furthermore, we demonstrate the reliable discrimination of the saccharide concentrations, as well as compositions in mixed solutions, which contain all five mono- and disaccharides simultaneously. These results underline the excellent discriminative power of plasmonic SEIRA spectroscopy in combination with the PCA. This unique combination and the insights gained will improve the detection of biomolecules in different complex environments.
In this paper, the problem of estimating the background of a spectrum is addressed. We propose to fit this background to a low-order polynomial, but rather than determining the polynomial parameters that minimise a least-squares criterion (i.e. a quadratic cost function), non-quadratic cost functions well adapted to the problem are proposed. To minimise these cost functions, we use the half-quadratic minimisation. It yields a fast and simple method, which can be applied to a wide range of spectroscopic signal. Guidelines for the choice of the design parameters are given and illustrated on simulated spectra. Finally, the effectiveness of the method is shown by processing experimental infrared and Raman spectra.
A technique entitled robust baseline estimation is introduced, which uses techniques of robust local regression to estimate baselines in spectra that consist of sharp features superimposed upon a continuous, slowly varying baseline. The technique is applied to synthetic spectra, to evaluate its capabilities, and to laser-induced fluorescence spectra of OH (produced from the reaction of ozone with hydrogen atoms). The latter example is a particularly challenging case for baseline estimation because the experimental noise varies as a function of frequency.
To increase the power and the robustness of spectroscopic process analyzers, methods are needed that suppress the spectral variation that is not related to the property of interest in the process stream. An approach for the selection of a suitable method is presented. The approach uses the net analyte signal (NAS) to analyze the situation and to select methods to suppress the nonrelevant spectral variation. The empirically determined signal-to-noise of the NAS is used as a figure of merit. The advantages of the approach are (i). that the error of the reference method does not affect method selection and (ii). that only a few spectral measurements are needed. A diagnostic plot is proposed that guides the user in the evaluation of the particular suppression method. As an example, NIR spectroscopic monitoring of a mol-sieve separation process is used.
- H F M Boelens
- P H C Eilers
Boelens, H.F.M. ; Eilers, P.H.C. ; Hankemeier Th. Anal. Chem. 2005, (in press)
- A Jirasek
- G Schulze
- M M L Yu
- M W Blades
- R F B Turner
Jirasek, A. ; Schulze G. ; Yu, M.M.L. ; Blades, M.W. ; Turner, R.F.B. Appl.
Spectrosc. 2004, 58, 1488-1499
- M Mazet
- C Carteret
- D Brie
- J Idier
- B Humbert
Mazet, M. ; Carteret, C. ; Brie, D. ; Idier, J. ; Humbert, B. Chemom. Intell. Lab.
Syst. 2005, 76, 121-133
- D A Mcnulty
- H J H Macfie
McNulty, D.A. ; MacFie, H.J.H. J. of Chemom. 1997, 11, 1-11
- P H C Eilers
- R X De Menezes
Eilers, P.H.C. ; de Menezes, R.X. Bioinformatics, 2005, 21, 1146-1153
- M A Knee
- H J Annegarn
Knee, M.A. and Annegarn H.J. Nucl. Instrum. Methods Phys. Rese. Sect. B, 1996,
109, 201-213
- A J Phillips
- P Hamilton
Phillips A.J. ; Hamilton P.A Anal. Chem. 1996, 68, 4020-4025
- B Bogaert
- H F M Boelens
- H C Smit
Bogaert, B. ; Boelens, H.F.M. ; Smit, H.C. Anal. Chim. Acta 1993, 274, 71-85
- P H C Eilers
Eilers, P.H.C. Anal. Chem. 2004, 76, 404-411
- V P Andreev
- T Rejtar
- H Chen
- E V Mosovets
- A R Ivanov
- B L Karger
Andreev, V.P. ; Rejtar, T. ; Chen, H. ; Mosovets, E.V. ; Ivanov A.R. ; Karger B.L.
Anal. Chem. 2003, 75, 6314-6326
- P H C Eilers
Eilers, P.H.C. Anal. Chem. 2003, 75, 3631-3636