Comparison of functional and discrete data analysis regimes for Raman spectra.

Raman spectral data are best described by mathematical functions; however, due to the spectroscopic measurement setup, only discrete points of these functions are measured. Therefore, we investigated the Raman spectral data for the first time in the functional framework. First, we approximated the Raman spectra by using B-spline basis functions. Afterwards, we applied the functional principal component analysis followed by the linear discriminant analysis (FPCA-LDA) and compared the results with those of the classical principal component analysis followed by the linear discriminant analysis (PCA-LDA). In this context, simulation and experimental Raman spectra were used. In the simulated Raman spectra, normal and abnormal spectra were used for a classification model, where the abnormal spectra were built by shifting one peak position. We showed that the mean sensitivities of the FPCA-LDA method were higher than the mean sensitivities of the PCA-LDA method, especially when the signal-to-noise ratio is low and the shift of the peak position is small. However, for a higher signal-to-noise ratio, both methods performed equally. Additionally, a slight improvement of the mean sensitivity could be shown if the FPCA-LDA method was applied to experimental Raman data.

Influence of Carbon Sources on Quantification of Deuterium Incorporation in Heterotrophic Bacteria: A Raman-Stable Isotope Labeling Approach.

A rapid and reliable method for the differentiation between active and inactive bacteria at single cell level is urgently needed in many fields including clinical diagnosis and environmental microbiology, to understand the contribution of metabolically active bacteria in fundamental processes triggering environmental and public health risks. Here, using heavy water (D2O) with Raman-stable isotope labeling (Raman-D2O), we evaluated the reliability of the quantification of deuterium uptake, a well-known indicator for the general metabolic activity of bacteria. For this purpose, we based our study on the quantification of deuterium assimilation from heavy water into single bacterial cells to check the influence of carbon source and bacterial identity on the deuterium uptake. We show that compared to complex carbon substrates, the deuterium assimilation is higher in the presence of simpler substrates such as sugars but differs significantly among bacterial isolates. Despite this variability, the developed classification models could differentiate deuterium labeled and nonlabeled single cells with high sensitivity and specificity. Highlighting the variability between single bacterial cells, the study emphasizes the challenges in establishing a threshold in terms of deuterium uptake to distinguish deuterium labeled and nonlabeled cells. Overall, we show that the Raman-D2O approach, when coupled with chemometrics, constitutes a powerful approach for monitoring single bacterial cells.

Deep learning as phase retrieval tool for CARS spectra.

Finding efficient and reliable methods for the extraction of the phase in optical measurements is challenging and has been widely investigated. Although sophisticated optical settings, e.g. holography, measure directly the phase, the use of algorithmic methods has gained attention due to its efficiency, fast calculation and easy setup requirements. We investigated three phase retrieval methods: the maximum entropy technique (MEM), the Kramers-Kronig relation (KK), and for the first time deep learning using the Long Short-Term Memory network (LSTM). LSTM shows superior results for the phase retrieval problem of coherent anti-Stokes Raman spectra in comparison to MEM and KK.

Comparison of denoising tools for the reconstruction of nonlinear multimodal images.

Biophotonic multimodal imaging techniques provide deep insights into biological samples such as cells or tissues. However, the measurement time increases dramatically when high-resolution multimodal images (MM) are required. To address this challenge, mathematical methods can be used to shorten the acquisition time for such high-quality images. In this research, we compared standard methods, e.g., the median filter method and the phase retrieval method via the Gerchberg-Saxton algorithm with artificial intelligence (AI) based methods using MM images of head and neck tissues. The AI methods include two approaches: the first one is a transfer learning-based technique that uses the pre-trained network DnCNN. The second approach is the training of networks using augmented head and neck MM images. In this manner, we compared the Noise2Noise network, the MIRNet network, and our deep learning network namely incSRCNN, which is derived from the super-resolution convolutional neural network and inspired by the inception network. These methods reconstruct improved images using measured low-quality (LQ) images, which were measured in approximately 2 seconds. The evaluation was performed on artificial LQ images generated by degrading high-quality (HQ) images measured in 8 seconds using Poisson noise. The results showed the potential of using deep learning on these multimodal images to improve the data quality and reduce the acquisition time. Our proposed network has the advantage of having a simple architecture compared with similar-performing but highly parametrized networks DnCNN, MIRNet, and Noise2Noise.

Bolstering fitness via CO2 fixation and organic carbon uptake: mixotrophs in modern groundwater.

Current understanding of organic carbon inputs into ecosystems lacking photosynthetic primary production is predicated on data and inferences derived almost entirely from metagenomic analyses. The elevated abundances of putative chemolithoautotrophs in groundwaters suggest that dark CO2 fixation is an integral component of subsurface trophic webs. To understand the impact of autotrophically fixed carbon, the flux of CO2-derived carbon through various populations of subsurface microbiota must first be resolved, both quantitatively and temporally. Here we implement novel Stable Isotope Cluster Analysis to render a time-resolved and quantitative evaluation of 13CO2-derived carbon flow through a groundwater community in microcosms stimulated with reduced sulfur compounds. We demonstrate that mixotrophs, not strict autotrophs, were the most abundant active organisms in groundwater microcosms. Species of Hydrogenophaga, Polaromonas, Dechloromonas, and other metabolically versatile mixotrophs drove the production and remineralization of organic carbon. Their activity facilitated the replacement of 43% and 80% of total microbial carbon stores in the groundwater microcosms with 13C in just 21 and 70 days, respectively. The mixotrophs employed different strategies for satisfying their carbon requirements by balancing CO2 fixation and uptake of available organic compounds. These different strategies might provide fitness under nutrient-limited conditions, explaining the great abundances of mixotrophs in other oligotrophic habitats, such as the upper ocean and boreal lakes.

Trends in artificial intelligence, machine learning, and chemometrics applied to chemical data.

Artificial intelligence-based methods such as chemometrics, machine learning, and deep learning are promising tools that lead to a clearer and better understanding of data. Only with these tools, data can be used to its full extent, and the gained knowledge on processes, interactions, and characteristics of the sample is maximized. Therefore, scientists are developing data science tools mentioned above to automatically and accurately extract information from data and increase the application possibilities of the respective data in various fields. Accordingly, AI-based techniques were utilized for chemical data since the 1970s and this review paper focuses on the recent trends of chemometrics, machine learning, and deep learning for chemical and spectroscopic data in 2020. In this regard, inverse modeling, preprocessing methods, and data modeling applied to spectra and image data for various measurement techniques are discussed.