FT-Raman and FTIR spectroscopy as a tools showing marker of platinum-resistant phenomena in women suffering from ovarian cancer.

Platinum-resistant phenomena in ovarian cancer is very dangerous for women suffering from this disease, because reduces the chances of complete recovery. Unfortunately, until now there are no methods to verify whether a woman with ovarian cancer is platinum-resistant. Importantly, histopathology images also were not shown differences in the ovarian cancer between platinum-resistant and platinum-sensitive tissues. Therefore, in this study, Fourier Transform InfraRed (FTIR) and FT-Raman spectroscopy techniques were used to find chemical differences between platinum-resistant and platinum-sensitive ovarian cancer tissues. Furthermore, Principal Component Analysis (PCA) and machine learning methods were performed to show if it possible to differentiate these two kind of tissues as well as to propose spectroscopy marker of platinum-resistant. Indeed, obtained results showed, that in platinum-resistant ovarian cancer tissues higher amount of phospholipids, proteins and lipids were visible, however when the ratio between intensities of peaks at 1637 cm-1 (FTIR) and at 2944 cm-1 (Raman) and every peaks in spectra was calculated, difference between groups of samples were not noticed. Moreover, structural changes visible as a shift of peaks were noticed for C-O-C, C-H bending and amide II bonds. PCA clearly showed, that PC1 can be used to differentiate platinum-resistant and platinum-sensitive ovarian cancer tissues, while two-trace two-dimensional correlation spectra (2T2D-COS) showed, that only in amide II, amide I and asymmetric CH lipids vibrations correlation between two analyzed types of tissues were noticed. Finally, machine learning algorithms showed, that values of accuracy, sensitivity and specificity were near to 100% for FTIR and around 95% for FT-Raman spectroscopy. Using decision tree peaks at 1777 cm-1, 2974 cm-1 (FTIR) and 1714 cm-1, 2817 cm-1 (FT-Raman) were proposed as spectroscopy marker of platinum-resistant.

FT-Raman spectra in combination with machine learning and multivariate analyses as a diagnostic tool in brain tumors.

Brain tumors are one of the most dangerous, because the position of these are in the organ that governs all life processes. Moreover, a lot of brain tumor types were observed, but only one main diagnostic method was used - histopathology, for which preparation of sample was long. Consequently, a new, quicker diagnostic method is needed. In this paper, FT-Raman spectra of brain tissues were analyzed by Principal Component Analysis (PCA), Hierarchical Cluster Analysis (HCA), four different machine learning (ML) algorithms to show possibility of differentiating between glioblastoma G4 and meningiomas, as well as two different types of meningiomas (atypical and angiomatous). Obtained results showed that in meningiomas additional peak around 1503 cm-1 and higher level of amides was noticed in comparison with glioblastoma G4. In the case of meningiomas differentiation, in angiomatous meningiomas tissues lower level of lipids and polysaccharides were visible than in atypical meningiomas. Moreover, PCA analyses showed higher distinction between glioblastoma G4 and meningiomas in the FT-Raman range between 800 cm-1 and 1800 cm-1 and between two types of meningiomas in the range between 2700 cm-1 and 3000 cm-1. Decision trees showed, that the most important peaks to differentiate glioblastoma and meningiomas were at 1151 cm-1 and 2836 cm-1 while for angiomatous and atypical meningiomas - 1514 cm-1 and 2875 cm-1. Furthermore, the accuracy of obtained results for glioblastoma G4 and meningiomas was 88 %, while for meningiomas - 92 %. Consequently, obtained data showed possibility of using FT-Raman spectroscopy in diagnosis of different types of brain tumors.

FT-Raman data analyzed by multivariate and machine learning as a new methods for detection spectroscopy marker of platinum-resistant women suffering from ovarian cancer.

The phenomenon of platinum resistance is a very serious problem in the treatment of ovarian cancer. Unfortunately, no molecular, genetic marker that could be used in assigning women suffering from ovarian cancer to the platinum-resistant or platinum-sensitive group has been discovered so far. Therefore, in this study, for the first time, we used FT-Raman spectroscopy to determine chemical differences and chemical markers presented in serum, which could be used to differentiate platinum-resistant and platinum-sensitive women. The result obtained showed that in the serum collected from platinum-resistant women, a significant increase of chemical compounds was observed in comparison with the serum collected from platinum-sensitive woman. Moreover, a decrease in the ratio between amides vibrations and shifts of peaks, respectively, corresponding to C-C/C-N stretching vibrations from proteins, amide III, amide II, C = O and CH lipids vibrations suggested that in these compounds, structural changes occurred. The Principal Component Analysis (PCA) showed that using FT-Raman range, where the above-mentioned functional groups were present, it was possible to differentiate the serum collected from both analyzed groups. Moreover, C5.0 decision tree clearly showed that Raman shifts at 1224 cm-1 and 2713 cm-1 could be used as a marker of platinum resistance. Importantly, machine learning methods showed that the accuracy, sensitivity and specificity of the FT-Raman spectroscopy were from 95 to 100%.

Application of the Fuzzy Approach for Evaluating and Selecting Relevant Objects, Features, and Their Ranges.

Relevant attribute selection in machine learning is a key aspect aimed at simplifying the problem, reducing its dimensionality, and consequently accelerating computation. This paper proposes new algorithms for selecting relevant features and evaluating and selecting a subset of relevant objects in a dataset. Both algorithms are mainly based on the use of a fuzzy approach. The research presented here yielded preliminary results of a new approach to the problem of selecting relevant attributes and objects and selecting appropriate ranges of their values. Detailed results obtained on the Sonar dataset show the positive effects of this approach. Moreover, the observed results may suggest the effectiveness of the proposed method in terms of identifying a subset of truly relevant attributes from among those identified by traditional feature selection methods.

An application of raman spectroscopy in combination with machine learning to determine gastric cancer spectroscopy marker.

BACKGROUND AND OBJECTIVE: Globally, gastric carcinoma (Gca) ranks fifth in terms of incidence and third in terms of mortality. Higher serum tumor markers (TMs) than those from healthy individuals, led to TMs clinical application as diagnostic biomarkers for Gca. Actually, there is no accurate blood test to diagnose Gca. METHODS: Raman spectroscopy is applied as an efficient, credible, minimally invasive technique to evaluate the serum TMs levels in blood samples. After curative gastrectomy, serum TMs levels are important in predicting the recurrence of gastric cancer, which must be detected early. The experimentally assesed TMs levels using Raman measurements and ELISA test were used to develop a prediction model based on machine learning techniques. A total of 70 participants diagnosed with gastric cancer after surgery (n = 26) and healthy (n = 44) were comrpised in this study. RESULTS: In the Raman spectra of gastric cancer patients, an additional peak at 1182 cm-1 was observed and, the Raman intensity of amide III, II, I, and CH2 proteins as well as lipids functional group was higher. Furthermore, Principal Component Analysis (PCA) showed, that it is possible to distinguish between the control and Gca groups using the Raman range between 800 and 1800 cm-1, as well as between 2700 and 3000 cm-1. The analysis of Raman spectra dynamics in gastric cancer and healthy patients showed, that the vibrations at 1302 and 1306 cm-1 were characteristic for cancer patients. In addition, the selected machine learning methods showed classification accuracy of more than 95%, while obtaining an AUROC of 0.98. Such results were obtained using Deep Neural Networks and the XGBoost algorithm. CONCLUSIONS: The obtained results suggest, that Raman shifts at 1302 and 1306 cm-1 could be spectroscopic markers of gastric cancer.

Fourier transform infrared spectroscopic marker of glioblastoma obtained from machine learning and changes in the spectra.

BACKGROUND: Glioblastoma is among the most malignant brain cancer with an average survival rate measured in months. In neurosurgical practice, it is considered impossible to completely remove a glioblastoma because of difficulties in the intraoperative assessment of the boundaries between healthy brain tissue and glioblastoma cells. Therefore, it is important to find a new, quick, cost-effective and useful neurosurgical practice method for the intraoperative differentiation of glioblastoma from healthy brain tissue. METHODS: Herein, the features of absorbance at specific wavenumbers considered characteristic of glioblastoma tissues could be markers of this cancer. We used Fourier transform infrared spectroscopy to measure the spectra of tissues collected from control and patients suffering from glioblastoma. RESULTS: The spectrum obtained from glioblastoma tissues demonstrated an additional peak at 1612 cm-1 and a shift of peaks at 1675 cm-1 and 1637 cm-1. Deconvolution of amide I vibrations showed that in the glioblastoma tissue, the percentage amount of ß-sheet is around 20% higher than that in the control. Moreover, the principal component analysis showed that using fingerprint and amide I regions it is possible to distinguish cancer and non-cancer samples. Machine learning methods presented that the accuracy of the results is around 100%. Finally, analysis of the differences in the rate of change of Fourier transform infrared spectroscopy spectra showed that absorbance features between 1053 cm-1 and 1056 cm-1 as well as between 1564 cm-1 and 1588 cm-1 are characteristic of glioblastoma. CONCLUSION: Calculated features of absorbance at specific wavenumbers could be used as a spectroscopic marker of glioblastoma which may be useful in the future for neuronavigation.

Correlation between human colon cancer specific antigens and Raman spectra. Attempting to use Raman spectroscopy in the determination of tumor markers for colon cancer.

Colorectal cancer is the second most common cause of cancer-related deaths worldwide. To follow up on the progression of the disease, tumor markers are commonly used. Here, we report serum analysis based on Raman spectroscopy to provide a rapid cancer diagnosis with tumor markers and two new cell adhesion molecules measured using the ELISA method. Raman spectra showed higher Raman intensities at 1447 cm-1 1560 cm-1, 1665 cm-1, and 1769 cm-1, which originated from CH2 proteins and lipids, amide II and amide I, and CO lipids vibrations. Furthermore, the correlation test showed, that only the CEA colon cancer marker correlated with the Raman spectra. Importantly, machine learning methods showed, that the accuracy of the Raman method in the detection of colon cancer was around 95 %. Obtained results suggest, that Raman shifts at 1302 cm-1 and 1306 cm-1 can be used as spectroscopy markers of colon cancer.

Increased levels of nerve growth factor accompany oxidative load in recurrent pregnancy loss. Machine learning applied to FT-Raman spectra study.

The presented article is focused on developing and validating an efficient, credible, minimally invasive technique based on spectral signatures of blood serum samples in patients with diagnosed recurrent pregnancy loss (RPL) versus healthy individuals who were followed at the Gynecology department. A total of 120 participants, RPL disease (n = 60) and healthy individuals (n = 60), participated in the study. First, we investigated the effect of circulating nerve growth factor (NGF) in RPL and healthy groups. To show NGF's effect, we measured the level of oxidative loads such as Total Antioxidant Level (TAS), Total Oxidant Level (TOS), and Oxidative Stress Index (OSI) with Beckman Coulter AU system and biochemical assays. We find a correlation between oxidative load and NGF level. Oxidative load mainly causes structural changes in the blood. Therefore, we obtained Raman measurements of the participant's serum. Then we selected two Raman regions, 800 and 1800 cm-1, and between 2700 cm-1 and 3000 cm-1, to see chemical changes. We noted that Raman spectra obtained for RPL and healthy women differed. The findings confirm that the imbalance between reactive oxygen species and antioxidants has important implications for the pathogenesis of RPL and that NGF levels accompany the level of oxidative load in the RPL state. Biomolecular structure and composition were determined using Raman spectroscopy and machine learning methods, and the correlation of these parameters was studied alongside machine learning technologies to advance toward clinical translation. Here we determined and validated the development of instrumentation for the Analysis of RPL patients' serum that can differentiate from control individuals with an accuracy of 100% using the Raman region corresponding to structural changes. Furthermore, this study found a correlation between traditional biochemical parameters and Raman data. This suggests that Raman spectroscopy is a sensitive tool for detecting biochemical changes in serum caused by RPL or other diseases.

Blood serum lipid profiling may improve the management of recurrent miscarriage: a combination of machine learning of mid-infrared spectra and biochemical assays.

The present article is focused on developing and validating an efficient, credible, minimally invasive technique based on spectral signatures of blood samples of women with recurrent miscarriage vs. those of healthy individuals who were followed in the Department of Obstetrics and Gynecology for 2 years. For this purpose, blood samples from a total of 120 participants, including healthy women (n=60) and women with diagnosed recurrent miscarriage (n=60), were obtained. The lipid profile (high-density lipoprotein, low-density lipoprotein, triglyceride, and total cholesterol levels) and lipid peroxidation (malondialdehyde and glutathione levels) were evaluated with a Beckman Coulter analyzer system for chemical analysis. Biomolecular structure and composition were determined using an attenuated total reflectance sampling methodology with Fourier transform infrared spectroscopy alongside machine learning technology to advance toward clinical translation. Here, we developed and validated instrumentation for the analysis of recurrent miscarriage patient serum that was able to differentiate recurrent miscarriage and control patients with an accuracy of 100% using a Fourier transform infrared region corresponding to lipids. We found that predictors of lipid profile abnormalities in maternal serum could significantly improve this patient pathway. The study also presents preliminary results from the first prospective clinical validation study of its kind.

COVID-19 antibody level analysis with feature selection approach.

The study presented here considers the analysis of a medical dataset for the identification of the stage of onset of COVID-19 coronavirus. These data, presented in previous work by the authors, have been subjected to extensive analysis and additional calculations. The data were obtained by analyzing blood samples of infected individuals at 1, 3, and 6 months after COVID-19 infection. Results were obtained from FTIR spectrometry experiments. The results indicate a very effective ability to identify the different states of infection, and between 1 and 6 months even perfect. Specific spectrometry wavelength ranges can also be distinguished as medical markers.

Apocynin reduces cytotoxic effects of monosodium glutamate in the brain: A spectroscopic, oxidative load, and machine learning study.

Herein, we examined the modulatory effects ofApocynum (APO) on Monosodium Glutamate (MSG)-induced oxidative damage on the brain tissue of rats after long-term consumption of blood serum components by biochemical assays, Fourier transform infrared spectroscopy(FTIR), and machine learning methods. Sprague-Dawley male rats were randomly divided into the Control, Control + APO, MSG, and MSG + APO groups (n = 8 per group). All administrations were made by oral gavage saline, MSG, or APO and they were repeated for 28 days of the experiments. Brain tissue and blood serum samples were collected and analyzed for measurement levels ofmalondialdehyde (MDA),glutathione (GSH),myeloperoxidase (MPO), superoxide dismutase (SOD) activity, and Spectroscopic analysis. After 29 days, the results were evaluated using machine learning (ML). The levels of MDA and MPO showed changes in the MSG and MSG + APO groups, respectively. Changes in the proteins and lipids were observed in the FTIR spectra of the MSG groups. Additionally, APO in these animals improved the FTIR spectra to be similar to those in the Control group. The accuracy of the FTIR results calculated by ML was 100%. The findings of this study demonstrate that Apocynin treatment protectsagainst MSG-induced oxidative damage by inhibitingreactive oxygen speciesand upregulatingantioxidant capacity, indicating its potential in alleviatingthe toxic effects of MSG.

Development of novel spectroscopic and machine learning methods for the measurement of periodic changes in COVID-19 antibody level.

In this research, blood samples of 47 patients infected by COVID were analyzed. The samples were taken on the 1st, 3rd and 6th month after the detection of COVID infection. Total antibody levels were measured against the SARS-CoV-2 N antigen and surrogate virus neutralization by serological methods. To differentiate COVID patients with different antibody levels, Fourier Transform InfraRed (FTIR) and Raman spectroscopy methods were used. The spectroscopy data were analyzed by multivariate analysis, machine learning and neural network methods. It was shown, that analysis of serum using the above-mentioned spectroscopy methods allows to differentiate antibody levels between 1 and 6 months via spectral biomarkers of amides II and I. Moreover, multivariate analysis showed, that using Raman spectroscopy in the range between 1317 cm-1 and 1432 cm-1, 2840 cm-1 and 2956 cm-1 it is possible to distinguish patients after 1, 3, and 6 months from COVID with a sensitivity close to 100%.

Determination of idiopathic female infertility from infrared spectra of follicle fluid combined with gonadotrophin levels, multivariate analysis and machine learning methods.

By in vitro fertilization, oocytes can be removed and the embryo can be cultured, and then trans cervically replaced when they reach cleavage or at the blastocyst stage. The characterization of the follicular fluid is important for the treatment process. Women who applied to the Academic Hospital in vitro fertilization (IVF) Center diagnosed with idiopathic female infertility (IFI) were sought in the patient group. Demographics and clinical gonadotropin measurements of the study population were recorded. Of the 116 follicular fluid samples (n=58 male-induced infertility; n=58 control) were analyzed using the FTIR system. To identify FTIR spectral characteristics of follicular fluids associated with an ovarian reserve and reproductive hormone levels from control and IFI, six machine learning methods and multivariate analysis were used. To assess the quantitative information about the total biochemical composition of a follicular fluid across various diagnoses. FTIR spectra showed a higher level of vibrations corresponding to lipids and a lower level of amide vibrations in the IFI group. Furthermore, the T square plot from Partial Last Square (PLS) analysis showed, that these vibrations can be used to distinguish IFI from the control group which was obtained by principal component analysis (PCA). Proteins and lipids play an important role in the development of IFI. The absorption dynamics of FTIR spectra showed wavenumbers with around 100% discrimination probability, which means, that the presented wavenumbers can be used as a spectroscopic marker of IFI. Also, six machine learning methods showed, that classification accuracy for the original set was from 93.75% to 100% depending on the learning algorithm used. These results can inform about IFI women's follicular fluid has biomacromolecular differentiation in their follicular fluid. By using a safe and effective tool for the characterization of changes in follicular fluid during in vitro fertilization, this study builds upon a comprehensive examination of the idiopathic female infertility remodeling process in human studies. We anticipate that this technology will be a valuable adjunct for clinical studies.

Correlation between endometriomas volume and Raman spectra. Attempting to use Raman spectroscopy in the diagnosis of endometrioma.

The formation of the uterus lining, i.e. the endometrium, outside the uterus (ex. in the abdominal cavity,ovaries,or anywhere in the body) is called endometriosis. The presence of endometrial tissue present in the ovaries, thickens after menstruation, leading to menstrual-like bleeding and to the formation of chocolate cyst (Endometrioma) because of the accumulation of old, brown blood in the ovary. It is still unknown, what triggers the development ofendometrioma. However,it leads to excessive bleeding during menstrual periods or abnormal bleeding between periods and infertility. Endometriosis is often first diagnosed in those who seek medical attention for infertility. Therefore, new markers of endometrioma as well as new methods of its diagnosis are sought. In this study we used Raman spectra of serum collected from 50 healthy women and 50 women suffering from endometriosis. The obtained Raman data were used in multivariateanalysis to determine the Raman range, which can be used for endometriomadiagnostics. Partial Least Square (PLS), Principal Component Analysis (PCA) and Hierarchical Component Analysis (HCA) showed, that it is possible to distinguish between the serum collected from healthy and un-healthy women using the Raman range between 800 cm-1 and 1800 cm-1 and between 2956 cm-1 and 2840 cm-1, while the first range corresponds to the fingerprint region and the second one to lipids vibrations. Consequently, the Pearson correlation test showeda significantpositive correlation betweenvaluesoflipidintensity in Raman spectra and volume of endometriomas. Summarizing, Raman spectroscopy can be a helpful tool in endometrioma diagnosis and the lipid vibrations are candidates for being a spectroscopic marker of the disease being studied.

Identification of polycystic ovary syndrome from blood serum using hormone levels via Raman spectroscopy and multivariate analysis.

Polycystic ovarian syndrome (PCOS) is a disease, which causes infertility in women. The factors for the development of the disease are still not well understood and diagnostic methods need to be improved. Therefore, in this study, Raman spectroscopy as a potential diagnostic tool, was investigated and spectra of blood serum were collected from PCOS and healthy women. The obtained spectra showed distinct changes in intensities as well as shift of peaks for the blood serum collected from PCOS compared to healthy individuals. Partial Last Square (PLS) analysis and Principal Component Analysis (PCA) allowed to determine that Raman shifts of amides (1500 - 1700 cm-1) and CH2, CH3 lipid groups (2700 - 3000 cm-1), could be thus used as potential PCOS markers. Furthermore, the Pearson correlation test showed a strong correlation between hormones (lutropin (LH), prolactin (PRL), follicle-stimulating (FSH), dehydroepiandrosterone (DHEAS), thyroid-stimulating (TSH), Estradiol) and lipids, as well as between hormones and protein functional groups in PCOS women, compared to the control. These results show, that the lipid and protein balance could be potentially applied as a helpful PCOS marker in Raman spectra.

Deep architectures for long-term stock price prediction with a heuristic-based strategy for trading simulations.

Stock price prediction is a popular yet challenging task and deep learning provides the means to conduct the mining for the different patterns that trigger its dynamic movement. In this paper, the task is to predict the close price for 25 companies enlisted at the Bucharest Stock Exchange, from a novel data set introduced herein. Towards this scope, two traditional deep learning architectures are designed in comparison: a long short-memory network and a temporal convolutional neural model. Based on their predictions, a trading strategy, whose decision to buy or sell depends on two different thresholds, is proposed. A hill climbing approach selects the optimal values for these parameters. The prediction of the two deep learning representatives used in the subsequent trading strategy leads to distinct facets of gain.

Active enhancer positions can be accurately predicted from chromatin marks and collective sequence motif data.

BACKGROUND: Transcriptional regulation in multi-cellular organisms is a complex process involving multiple modular regulatory elements for each gene. Building whole-genome models of transcriptional networks requires mapping all relevant enhancers and then linking them to target genes. Previous methods of enhancer identification based either on sequence information or on epigenetic marks have different limitations stemming from incompleteness of each of these datasets taken separately. RESULTS: In this work we present a new approach for discovery of regulatory elements based on the combination of sequence motifs and epigenetic marks measured with ChIP-Seq. Our method uses supervised learning approaches to train a model describing the dependence of enhancer activity on sequence features and histone marks. Our results indicate that using combination of features provides superior results to previous approaches based on either one of the datasets. While histone modifications remain the dominant feature for accurate predictions, the models based on sequence motifs have advantages in their general applicability to different tissues. Additionally, we assess the relevance of different sequence motifs in prediction accuracy showing that even tissue-specific enhancer activity depends on multiple motifs. CONCLUSIONS: Based on our results, we conclude that it is worthwhile to include sequence motif data into computational approaches to active enhancer prediction and also that classifiers trained on a specific set of enhancers can generalize with significant accuracy beyond the training set.