An infrared spectral signature of human lymphocyte subpopulations from peripheral blood.

Metastatic melanomas are frequently refractory to most adjuvant therapies such as chemotherapies and radiotherapies. Recently, immunotherapies have shown good results in the treatment of some metastatic melanomas. Immune cell infiltration in the tumor has been associated with successful immunotherapy. More generally, tumor infiltrating lymphocytes (TILs) in the primary tumor and in metastases of melanoma patients have been demonstrated to correlate positively with favorable clinical outcomes. Altogether, these findings suggest the importance of being able to identify, quantify and characterize immune infiltration at the tumor site for a better diagnostic and treatment choice. In this paper, we used Fourier Transform Infrared (FTIR) imaging to identify and quantify different subpopulations of T cells: the cytotoxic T cells (CD8+), the helper T cells (CD4+) and the regulatory T cells (T reg). As a proof of concept, we investigated pure populations isolated from human peripheral blood from 6 healthy donors. These subpopulations were isolated from blood samples by magnetic labeling and purities were assessed by Fluorescence Activated Cell Sorting (FACS). The results presented here show that Fourier Transform Infrared (FTIR) imaging followed by supervised Partial Least Square Discriminant Analysis (PLS-DA) allows an accurate identification of CD4+ T cells and CD8+ T cells (>86%). We then developed a PLS regression allowing the quantification of T reg in a different mix of immune cells (e.g. Peripheral Blood Mononuclear Cells (PBMCs)). Altogether, these results demonstrate the sensitivity of infrared imaging to detect the low biological variability observed in T cell subpopulations.

Preferred Spatial Frequencies for Human Face Processing Are Associated with Optimal Class Discrimination in the Machine

Psychophysical studies suggest that humans preferentially use a narrow band of low spatial frequencies for face recognition. Here we asked whether artificial face recognition systems have an improved recognition performance at the same spatial frequencies as humans. To this end, we estimated recognition performance over a large database of face images by computing three discriminability measures: Fisher Linear Discriminant Analysis, Non-Parametric Discriminant Analysis, and Mutual Information. In order to address frequency dependence, discriminabilities were measured as a function of (filtered) image size. All three measures revealed a maximum at the same image sizes, where the spatial frequency content corresponds to the psychophysical found frequencies. Our results therefore support the notion that the critical band of spatial frequencies for face recognition in humans and machines follows from inherent properties of face images, and that the use of these frequencies is associated with optimal face recognition performance.

Intelligent Internet Data Searching

In this thesis author approaches the problem of automated text classification, which is one of basic tasks for building Intelligent Internet Search Agent. The work discusses various approaches to solving sub-problems of automated text classification, such as feature extraction and machine learning on text sources. Author also describes her own multiword approach to feature extraction and pres-ents the results of testing this approach using linear discriminant analysis based classifier, and classifier combining unsupervised learning for etalon extraction with supervised learning using common backpropagation algorithm for multilevel perceptron.

Lajinvaihtoaikojen minimointi pakastetehtaassa

Diplomityön tarkoituksena oli selvittää miten lajinvaihtoaikoja voidaan vähentää ryhmittäin pakasteleipomossa. Työn osatavoitteina oli jakaa tuotteet ryhmiin sekä selvittää todellinen vaihtoaika kuuden kuukauden ajalta, jolloin saatiin työhön tarvittava vertailuaineisto. Työ rajattiin koskemaan vain yrityksen tehokkainta linjaa, koska siinä valmistetaan eniten tuotteita. Linjan tuotteet jaettiin ryhmiin erilaisten ominaisuuksien perusteella. Vaihtoaikojen lyhennyksessä sovellettiin eri teorioita. Tärkeimpinä teorioina voidaan mainita Shigeo Shingon kehittämä SMED-menetelmä, 5S-prosessi ja ryhmäanalyysi. SMED-menetelmän tavoitteena on jakaa asetukset sisäisiin ja ulkoisiin asetuksiin ja erottaa ne toisistaan. Tavoitteena on myös siirtää sisäisiä asetuksia ulkoisiksi. 5S-prosessi on visuaalista johtamista, jonka tavoitteena on pitää työympäristö siistinä. Ryhmäanalyysissä tuotteet jaetaan ensin ryhmiin j a sen jälkeen tuotteet laitetaan ryhmien sisällä parhaaseen mahdolliseen ajojärjestykseen. Tämän jälkeen ryhmät laitetaan keskenään parhaaseen ajojärjestykseen. Työn tavoitteena oli vähentää vaihtoaikaa viisi prosenttia tuotannon kokonaisajasta sekä tehdä kehityssuunnitelma, jonka avulla voidaan vähentää vaihtoaikoja kohdeyrityksen muilla linjoilla. Kokeilujen jälkeen kohdelinjan keskimääräinen viikoittainen vaihtoaika lyheni 1,1 % ja keskimääräisen vaihdon pituus lyheni 19 minuuttia. Tulosten perusteella kehitettiin kahdeksankohtainen kehityssuunnitelma.

Diagnosis of Alzheimer’s Disease from EEG by Means of Synchrony Measures in Optimized Frequency Bands

Several clinical studies have reported that EEG synchrony is affected by Alzheimer’s disease (AD). In this paper a frequency band analysis of AD EEG signals is presented, with the aim of improving the diagnosis of AD using EEG signals. In this paper, multiple synchrony measures are assessed through statistical tests (Mann–Whitney U test), including correlation, phase synchrony and Granger causality measures. Moreover, linear discriminant analysis (LDA) is conducted with those synchrony measures as features. For the data set at hand, the frequency range (5-6Hz) yields the best accuracy for diagnosing AD, which lies within the classical theta band (4-8Hz). The corresponding classification error is 4.88% for directed transfer function (DTF) Granger causality measure. Interestingly, results show that EEG of AD patients is more synchronous than in healthy subjects within the optimized range 5-6Hz, which is in sharp contrast with the loss of synchrony in AD EEG reported in many earlier studies. This new finding may provide new insights about the neurophysiology of AD. Additional testing on larger AD datasets is required to verify the effectiveness of the proposed approach.

Tracking Invasion Histories in the Sea: Facing Complex Scenarios Using Multilocus Data

In recent years, new analytical tools have allowed researchers to extract historical information contained in molecular data, which has fundamentally transformed our understanding of processes ruling biological invasions. However, the use of these new analytical tools has been largely restricted to studies of terrestrial organisms despite the growing recognition that the sea contains ecosystems that are amongst the most heavily affected by biological invasions, and that marine invasion histories are often remarkably complex. Here, we studied the routes of invasion and colonisation histories of an invasive marine invertebrate Microcosmus squamiger (Ascidiacea) using microsatellite loci, mitochondrial DNA sequence data and 11 worldwide populations. Discriminant analysis of principal components, clustering methods and approximate Bayesian computation (ABC) methods showed that the most likely source of the introduced populations was a single admixture event that involved populations from two genetically differentiated ancestral regions - the western and eastern coasts of Australia. The ABC analyses revealed that colonisation of the introduced range of M. squamiger consisted of a series of non-independent introductions along the coastlines of Africa, North America and Europe. Furthermore, we inferred that the sequence of colonisation across continents was in line with historical taxonomic records - first the Mediterranean Sea and South Africa from an unsampled ancestral population, followed by sequential introductions in California and, more recently, the NE Atlantic Ocean. We revealed the most likely invasion history for world populations of M. squamiger, which is broadly characterized by the presence of multiple ancestral sources and non-independent introductions within the introduced range. The results presented here illustrate the complexity of marine invasion routes and identify a cause-effect relationship between human-mediated transport and the success of widespread marine non-indigenous species, which benefit from stepping-stone invasions and admixture processes involving different sources for the spread and expansion of their range.

Distribuição de metais pesados em sedimentos do sistema estuarino da Ilha de Vitória-ES

Heavy-metal concentration in sediment is an important parameter for identifying pollution sources and assessing pollution levels in aquatic ecosystems. In this context, the present study aimed at determining concentrations of heavy metals in sediments from the Vitória estuarine system, Brazil. Twenty nine stations were surveyed to assess the spatial distribution of heavy metals. The metals for silt-clay fractions (<63 µm) were analyzed through atomic absorption spectrometry. A discriminant analysis segregated the stations in four groups representing four areas within the estuarine system. The Espírito Santo Bay showed the lowest metal concentrations, while the Vitória harbor canal showed the highest. We concluded that concentrations of heavy metals reflect natural conditions and the contribution of human activities from sewage and industrial effluents. It was not possible to directly associate metal concentrations to specific pollution sources.

Avaliação da concentração de alguns íons metálicos em diferentes espécies de líquens do cerrado Sul-Mato-Grossense

Concentrations of Fe, Mn, Co, Cr, Zn and Cu were determinated using flame atomic absorption spectrometry in nine lichen species of the Sul-Mato-Grossense cerrado. The average metal ion concentrations varied in the following ranges: Fe, 248.41-1568.01; Mn, 98.50-397.33; Co, 10.08-24.81; Cr, 18.24-44.26; Zn, 14.62-34.79 and Cu, 3.23-7.57 mg kg-1. Statistical analysis (Pearson and Cluster) applied to the metal ion concentrations indicated that the accumulation of these ions can be due to several anthropogenic sources including agricultural activities, mineral exploration, biomass burning, soil mineral composition and leather tanning processes by chromium.

Distinção entre cachaças destiladas em alambiques e em colunas usando quimiometria

One hundred fifteen cachaça samples derived from distillation in copper stills (73) or in stainless steels (42) were analyzed for thirty five itens by chromatography and inductively coupled plasma optical emission spectrometry. The analytical data were treated through Factor Analysis (FA), Partial Least Square Discriminant Analysis (PLS-DA) and Quadratic Discriminant Analysis (QDA). The FA explained 66.0% of the database variance. PLS-DA showed that it is possible to distinguish between the two groups of cachaças with 52.8% of the database variance. QDA was used to build up a classification model using acetaldehyde, ethyl carbamate, isobutyl alcohol, benzaldehyde, acetic acid and formaldehyde as chemical descriptors. The model presented 91.7% of accuracy on predicting the apparatus in which unknown samples were distilled.

Hidrocarbonetos policíclicos aromáticos (HPAS) em cachaça, rum, uísque e álcool combustível

The concentration of 15 polycyclic aromatic hydrocarbons (PAHs) in 57 samples of distillates (cachaça, rum, whiskey, and alcohol fuel) has been determined by HPLC-Fluorescence detection. The quantitative analytical profile of PAHs treated by Partial Least Square - Discriminant Analysis (PLS-DA) provided a good classification of the studied spirits based on their PAHs content. Additionally, the classification of the sugar cane derivatives according to the harvest practice was obtained treating the analytical data by Linear Discriminant Analysis (LDA), using naphthalene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benz[a]anthracene, benz[b]fluoranthene, and benz[g,h,i]perylene, as a chemical descriptors.

Aplicación de una lengua electrónica voltamétrica para la clasificación de vinos y estudio de correlación con la caracterización química y sensorial

This paper presents the analytical application of a novel electronic tongue based on voltammetric sensors array. This device was used in the classification of wines aged in barrels of different origins and toasting levels. Furthermore, a study of correlation between the response of the electronic tongue and the sensory and chemical characterization of samples was carried out. The results were evaluated by applying both principal component analysis and cluster analysis. The samples were clearly classified. Their distribution showed a high correspondence degree with the characteristics of the analyzed wines, it also showed similarity with the classification obtained from organoleptic analysis.

Composição química da precipitação úmida da região metropolitana de Porto Alegre, Brasil, 2005- 2007

This work aims to quantify the wet precipitation the Metropolitan Area of Porto Alegre (MAPA), in southern Brazil, through the analysis of major ions (by ion chromatography) and metallic elements (ICP/AES). By principal components analysis and cluster analysis was possible to identify the influence of natural and anthropic sources in wet precipitation. The results indicated of the higher contribution to the ions NH4+, SO4(2-) and Ca2+. Thus it was possible to identify the contribution of anthropogenic sources in wet precipitation in the study area, such as power plants, oil refineries, steel and vehicle emissions.

DISCRIMINAÇÃO DE MADEIRAS SIMILARES POR FLUORESCÊNCIA MOLECULAR E MÍNIMOS QUADRADOS PARCIAIS

Wood is an extremely complex biological material, which can show macroscopic similarities that make it difficult to discriminate between species. Discrimination between similar wood species can be achieved by either anatomic or instrumental methods, such as near infrared spectroscopy (NIR). Although different spectroscopy methods are currently available, few studies have applied them to discriminate between wood species. In this study, we applied a partial least squares-discriminant analysis (PLS-DA) model to evaluate the viability of using direct fluorescence measurements for discriminating between Eucalyptus grandis, Eucalyptus urograndis, and Cedrela odorata. The results show that molecular fluorescence is an efficient technique for discriminating between these visually similar wood species. With respect to calibration and the validation samples, we observed no misclassifications or outliers.

Mathematical modeling of level of anaesthesia from EEG measurements

Problem of modeling of anaesthesia depth level is studied in this Master Thesis. It applies analysis of EEG signals with nonlinear dynamics theory and further classification of obtained values. The main stages of this study are the following: data preprocessing; calculation of optimal embedding parameters for phase space reconstruction; obtaining reconstructed phase portraits of each EEG signal; formation of the feature set to characterise obtained phase portraits; classification of four different anaesthesia levels basing on previously estimated features. Classification was performed with: Linear and quadratic Discriminant Analysis, k Nearest Neighbours method and online clustering. In addition, this work provides overview of existing approaches to anaesthesia depth monitoring, description of basic concepts of nonlinear dynamics theory used in this Master Thesis and comparative analysis of several different classification methods.

Predicting bankruptcy of Finnish limited liability companies from historical financial statements

This master’s thesis studies the probability of bankruptcy of Finnish limited liability companies as a part of credit risk assessment. The main idea of this thesis is to build and test bankruptcy prediction models for Finnish limited liability companies that can be utilized in credit decision making. The data used in this thesis consists of historical financial statements from 2112 Finnish limited liability companies, half of which have filed for bankruptcy. A total of four models are developed, two with logistic regression and two with multivariate discriminant analysis (MDA). The time horizon of the models varies from 1 to 2 years prior to the bankruptcy, and 14 different financial variables are used in the model formation. The results show that the prediction accuracy of the models ranges between 81.7% and 88.9%, and the best prediction accuracy is achieved with the one year prior the bankruptcy logistic regression model. However the difference between the best logistic model and the best MDA model is minimal. Overall based on the results of this thesis it can be concluded that predicting bankruptcy is possible to some extent, but naturally the results are not perfect.