Learning Preferences with Kernel-Based Methods

Learning of preference relations has recently received significant attention in machine learning community. It is closely related to the classification and regression analysis and can be reduced to these tasks. However, preference learning involves prediction of ordering of the data points rather than prediction of a single numerical value as in case of regression or a class label as in case of classification. Therefore, studying preference relations within a separate framework facilitates not only better theoretical understanding of the problem, but also motivates development of the efficient algorithms for the task. Preference learning has many applications in domains such as information retrieval, bioinformatics, natural language processing, etc. For example, algorithms that learn to rank are frequently used in search engines for ordering documents retrieved by the query. Preference learning methods have been also applied to collaborative filtering problems for predicting individual customer choices from the vast amount of user generated feedback. In this thesis we propose several algorithms for learning preference relations. These algorithms stem from well founded and robust class of regularized least-squares methods and have many attractive computational properties. In order to improve the performance of our methods, we introduce several non-linear kernel functions. Thus, contribution of this thesis is twofold: kernel functions for structured data that are used to take advantage of various non-vectorial data representations and the preference learning algorithms that are suitable for different tasks, namely efficient learning of preference relations, learning with large amount of training data, and semi-supervised preference learning. Proposed kernel-based algorithms and kernels are applied to the parse ranking task in natural language processing, document ranking in information retrieval, and remote homology detection in bioinformatics domain. Training of kernel-based ranking algorithms can be infeasible when the size of the training set is large. This problem is addressed by proposing a preference learning algorithm whose computation complexity scales linearly with the number of training data points. We also introduce sparse approximation of the algorithm that can be efficiently trained with large amount of data. For situations when small amount of labeled data but a large amount of unlabeled data is available, we propose a co-regularized preference learning algorithm. To conclude, the methods presented in this thesis address not only the problem of the efficient training of the algorithms but also fast regularization parameter selection, multiple output prediction, and cross-validation. Furthermore, proposed algorithms lead to notably better performance in many preference learning tasks considered.

Decay of linkage disequilibrium within genes across HGDP-CEPH human samples: most population isolates do not show increased LD

Background It is well known that the pattern of linkage disequilibrium varies between human populations, with remarkable geographical stratification. Indirect association studies routinely exploit linkage disequilibrium around genes, particularly in isolated populations where it is assumed to be higher. Here, we explore both the amount and the decay of linkage disequilibrium with physical distance along 211 gene regions, most of them related to complex diseases, across 39 HGDP-CEPH population samples, focusing particularly on the populations defined as isolates. Within each gene region and population we use r2 between all possible single nucleotide polymorphism (SNP) pairs as a measure of linkage disequilibrium and focus on the proportion of SNP pairs with r2 greater than 0.8. Results Although the average r2 was found to be significantly different both between and within continental regions, a much higher proportion of r2 variance could be attributed to differences between continental regions (2.8% vs. 0.5%, respectively). Similarly, while the proportion of SNP pairs with r2 > 0.8 was significantly different across continents for all distance classes, it was generally much more homogenous within continents, except in the case of Africa and the Americas. The only isolated populations with consistently higher LD in all distance classes with respect to their continent are the Kalash (Central South Asia) and the Surui (America). Moreover, isolated populations showed only slightly higher proportions of SNP pairs with r2 > 0.8 per gene region than non-isolated populations in the same continent. Thus, the number of SNPs in isolated populations that need to be genotyped may be only slightly less than in non-isolates. Conclusion The 'isolated population' label by itself does not guarantee a greater genotyping efficiency in association studies, and properties other than increased linkage disequilibrium may make these populations interesting in genetic epidemiology.

Design and production of protein nanostructures for biomolecular detection

Particulate nanostructures are increasingly used for analytical purposes. Such particles are often generated by chemical synthesis from non-renewable raw materials. Generation of uniform nanoscale particles is challenging and particle surfaces must be modified to make the particles biocompatible and water-soluble. Usually nanoparticles are functionalized with binding molecules (e.g., antibodies or their fragments) and a label substance (if needed). Overall, producing nanoparticles for use in bioaffinity assays is a multistep process requiring several manufacturing and purification steps. This study describes a biological method of generating functionalized protein-based nanoparticles with specific binding activity on the particle surface and label activity inside the particles. Traditional chemical bioconjugation of the particle and specific binding molecules is replaced with genetic fusion of the binding molecule gene and particle backbone gene. The entity of the particle shell and binding moieties are synthesized from generic raw materials by bacteria, and fermentation is combined with a simple purification method based on inclusion bodies. The label activity is introduced during the purification. The process results in particles that are ready-to-use as reagents in bioaffinity. Apoferritin was used as particle body and the system was demonstrated using three different binding moieties: a small protein, a peptide and a single chain Fv antibody fragment that represents a complex protein including disulfide bridge.If needed, Eu3+ was used as label substance. The results showed that production system resulted in pure protein preparations, and the particles were of homogeneous size when visualized with transmission electron microscopy. Passively introduced label was stably associated with the particles, and binding molecules genetically fused to the particle specifically bound target molecules. Functionality of the particles in bioaffinity assays were successfully demonstrated with two types of assays; as labels and in particle-enhanced agglutination assay. This biological production procedure features many advantages that make the process especially suited for applications that have frequent and recurring requirements for homogeneous functional particles. The production process of ready, functional and watersoluble particles follows principles of “green chemistry”, is upscalable, fast and cost-effective.

Potencial de biocatálise enantiosseletiva de lipases microbianas

Microbial lipases have a great potential for commercial applications due to their stability, selectivity and broad substrate specificity because many non-natural acids, alcohols or amines can be used as the substrate. Three microbial lipases isolated from Brazilian soil samples (Aspergillus niger; Geotrichum candidum; Penicillium solitum) were compared in terms of their stability and as biocatalysts in the enantioselective esterification using racemic substrates in organic medium. The lipase from Aspergillus niger showed the highest activity (18.2 U/mL) and was highly thermostable, retaining 90% and 60% activity at 50 ºC and 60 ºC after 1 hour, respectively. In organic medium, this lipase provided the best results in terms of enantiomeric excess of the (S)-active acid (ee = 6.1%) and conversion value (c = 20%) in the esterification of (R,S)-ibuprofen with 1-propanol in isooctane. The esterification reaction of the racemic mixture of (R,S)-2-octanol with decanoic acid proceeded with high enantioselectivity when lipase from Aspergillus niger (E = 13.2) and commercial lipase from Candida antarctica (E = 20) were employed.

Supporting older people’s independent living at home through social and health care collaboration

Tässä tutkimuksessa tarkasteltiin ikäihmisten kotona asumista sosiaali- ja terveydenhuollon yhteistyön näkökulmasta. Tutkimuksen tarkoituksena oli lisätä ymmärrystä iäkkäiden kotihoidon asiakkaiden voimavaroista arjesta selviytymisen näkökulmasta, ja tutkia miten asiakkaiden hoito sosiaali- ja terveydenhuollon yhteistyönä toteutuu. Tutkimus oli poikkileikkaustutkimus, jossa sovellettiin kuvailevaa ja vertailevaa tutkimusasetelmaa. Tutkimusaineisto kerättiin yhden länsisuomalaisen kunnan kotihoidon asiakkailta (≥65 v.) ja heitä hoitavilta ammattihenkilöiltä. Kotihoidon 21 iäkästä asiakasta kuvasivat omia voimavarojaan arjesta selviytymisen näkökulmasta sekä kokemuksiaan hoidon toteutumisesta ammattihenkilöiden yhteistyönä. Aineisto kerättiin avoimella haastattelulla ja analysoitiin sisällön analyysillä. Lisäksi 25 kotihoidon ammattihenkilöä: 13 kotipalvelun työntekijää, 11 kotisairaanhoitajaa ja lääkäri kuvasivat kokemuksiaan iäkkään asiakkaan hoidon toteutumisesta ammattihenkilöiden yhteistyönä. Aineisto kerättiin fokusryhmähaastattelulla ja analysoitiin sisällön analyysillä. Näiden tulosten sekä aikaisemman kirjallisuuden perusteella laadittiin strukturoitu kyselylomake, jolla analysoitiin ja vertailtiin asiakkaiden ja ammattihenkilöiden näkemyksiä siitä, miten asiakkaiden hoito sosiaali- ja terveydenhuollon yhteistyönä toteutui. Esitestausten jälkeen kyselylomake lähetettiin 200 kotihoidon asiakkaalle ja 570 heitä hoitavalle kotihoidon työntekijälle: 485 kotipalvelun työntekijälle, 81 kotisairaanhoitajalle ja 4 lääkärille. Kyselyyn vastasi 120 asiakasta (60 %) ja 370 ammattihenkilöä (65 %). Ryhmien välisten erojen tarkastelussa käytettiin ristiintaulukointia, Pearsonin khin neliötestiä ja Fisherin tarkan todennäköisyyden testiä. Iäkkäiden asiakkaiden kuvauksissa voimavarat muodostuivat elämänhallinnan tunteesta ja toimintatahdon säilymisestä. Asiakkaat ammensivat arkeen voimaa harrastuksista ja sosiaalisesta verkostosta, mutta ulkopuolisten asettamat elämisen ehdot, terveydentilan heikkeneminen sekä yksinäisyys asettivat ikäihmisen ja hänen voimavaransa suurten haasteiden eteen. Tulokset osoittivat, että ammattihenkilöiden toiminta oli osittain ristiriidassa ikäihmisten omien odotusten kanssa, eikä se kaikilta osin tukenut asiakkaiden omia voimavaroja. Ammattihenkilöt tekivät hoitoon liittyviä päätöksiä ja toimintoja asiakkaiden puolesta, vaikka asiakkaille itselleen oli tärkeää elämänhallinnan tunne ja toimintatahdon säilyminen. Asiakkaiden voimavarojen tukemista moniammatillisena yhteistyönä vaikeuttivat ammattihenkilöiden vaikeus tunnistaa asiakkaiden omia voimavaroja sekä niitä uhkaavia tekijöitä, tiedon kulun ongelmat, tavoitteeton ja epäyhtenäinen tapa toimia sekä ammattihenkilöiden vastakkain asettuvat näkemyserot ja toimintatavat. Asiakkaiden ja ammattihenkilöiden näkemykset toteutetusta hoidosta erosivat toisistaan tilastollisesti merkitsevästi (p<0.05). Asiakkaat arvioivat sekä itsenäiseen toimintaan tukemisen että fyysisen, psyykkisen ja sosiaalisen tuen toteutuneen työntekijöitä huonommin. Yhteistyön kehittämishaasteita kotihoidossa ovat asiakkaan oman elämänsä asiantuntijuuden vahvistaminen, toimintakulttuurin muuttaminen asiakaslähtöiseksi tavoitteelliseksi toiminnaksi, ammattihenkilöiden roolien ja vastuun selkiyttäminen sekä tiedon kulun menetelmien kehittäminen. Tutkimus vahvistaa gerontologisen hoitotieteen tietoperustaa ja tuottaa uutta tietoa, jota voidaan soveltaa sosiaali- ja terveysalan koulutuksessa ja johtamisessa

Determinação espectrofotométrica simultânea de paracetamol e ibuprofeno em formulações farmacêuticas usando calibração multivariada

A simple method was proposed for determination of paracetamol and ibuprofen in tablets, based on UV measurements and partial least squares. The procedure was performed at pH 10.5, in the concentration ranges 3.00-15.00 µg ml-1 (paracetamol) and 2.40-12.00 µg ml-1 (ibuprofen). The model was able to predict paracetamol and ibuprofen in synthetic mixtures with root mean squares errors of prediction of 0.12 and 0.17 µg ml-1, respectively. Figures of merit (sensitivity, limit of detection and precision) were also estimated. The results achieved for the determination of these drugs in pharmaceutical formulations were in agreement with label claims and verified by HPLC.

Redução de vitamina C em suco de caju (Anacardium occidentale L.) industrializado e cajuína

Vitamin C degradation was evaluated in industrialized cashew juice of high pulp content and in cajuina by the method of Tillmans during eleven days of storage after the opening of the flask. For recently opened juices, vitamin C was found in the concentration range of 112 to 170 mg for 100 g of juice. The degradation of vitamin C in industrialized cashew juices changes when different additives are used. All of the cajuinas presented a vitamin C content below that specified on the label.

Long-term functional improvements in the 2-year treatment of schizophrenia outpatients with olanzapine long-acting injection

BACKGROUND: Little is known about the long-term changes in the functioning of schizophrenia patients receiving maintenance therapy with olanzapine long-acting injection (LAI), and whether observed changes differ from those seen with oral olanzapine. METHODS: This study describes changes in the levels of functioning among outpatients with schizophrenia treated with olanzapine-LAI compared with oral olanzapine over 2 years. This was a secondary analysis of data from a multicenter, randomized, open-label, 2-year study comparing the long-term treatment effectiveness of monthly olanzapine-LAI (405 mg/4 weeks; n=264) with daily oral olanzapine (10 mg/day; n=260). Levels of functioning were assessed with the Heinrichs-Carpenter Quality of Life Scale. Functional status was also classified as 'good', 'moderate', or 'poor', using a previous data-driven approach. Changes in functional levels were assessed with McNemar's test and comparisons between olanzapine-LAI and oral olanzapine employed the Student's t-test. RESULTS: Over the 2-year study, the patients treated with olanzapine-LAI improved their level of functioning (per Quality of Life total score) from 64.0-70.8 (P<0.001). Patients on oral olanzapine also increased their level of functioning from 62.1-70.1 (P<0.001). At baseline, 19.2% of the olanzapine-LAI-treated patients had a 'good' level of functioning, which increased to 27.5% (P<0.05). The figures for oral olanzapine were 14.2% and 24.5%, respectively (P<0.001). Results did not significantly differ between olanzapine-LAI and oral olanzapine. CONCLUSION: In this 2-year, open-label, randomized study of olanzapine-LAI, outpatients with schizophrenia maintained or improved their favorable baseline level of functioning over time. Results did not significantly differ between olanzapine-LAI and oral olanzapine.

Validação de métodos cromatográficos por clae para análise das vitaminas B1, B2, B6 e niacina naturalmente presentes em farinha de cereais

Complex B vitamins are present in some cereal foods and the ingestion of enriched products contributes to the recommended dietary intake of these micronutrients. To adapt the label of some products, it is necessary to develop and validate the analytical methods. These methods must be reliable and with enough sensitivity to analyze complex B vitamins naturally present in food at low concentration. The purpose of this work is to evaluate, with validated methods, the content of vitamins B1, B2, B6 and niacin in five cereal flours used in food industry (oat, rice, barley, corn and wheat).

Machine Learning and Clinical Text. Supporting Health Information Flow

Fluent health information flow is critical for clinical decision-making. However, a considerable part of this information is free-form text and inabilities to utilize it create risks to patient safety and cost-­effective hospital administration. Methods for automated processing of clinical text are emerging. The aim in this doctoral dissertation is to study machine learning and clinical text in order to support health information flow.First, by analyzing the content of authentic patient records, the aim is to specify clinical needs in order to guide the development of machine learning applications.The contributions are a model of the ideal information flow,a model of the problems and challenges in reality, and a road map for the technology development. Second, by developing applications for practical cases,the aim is to concretize ways to support health information flow. Altogether five machine learning applications for three practical cases are described: The first two applications are binary classification and regression related to the practical case of topic labeling and relevance ranking.The third and fourth application are supervised and unsupervised multi-class classification for the practical case of topic segmentation and labeling.These four applications are tested with Finnish intensive care patient records.The fifth application is multi-label classification for the practical task of diagnosis coding. It is tested with English radiology reports.The performance of all these applications is promising. Third, the aim is to study how the quality of machine learning applications can be reliably evaluated.The associations between performance evaluation measures and methods are addressed,and a new hold-out method is introduced.This method contributes not only to processing time but also to the evaluation diversity and quality. The main conclusion is that developing machine learning applications for text requires interdisciplinary, international collaboration. Practical cases are very different, and hence the development must begin from genuine user needs and domain expertise. The technological expertise must cover linguistics,machine learning, and information systems. Finally, the methods must be evaluated both statistically and through authentic user-feedback.

Elementos-traço e sódio em suco de uva: aspectos nutricionais e toxicológicos

Eight trace elements were determined in 20 Brazilian brands of grape juice, distributed over the country. Highest measured concentrations (As: 0.016; Cd: 0.010; Cr: 0.060; Cu: 1.28; Ni: 0.032; Pb: 0.016; Sb: 0.0040 and Zn: 1.44 mg L-1) comply with Brazilian maximal tolerance levels for inorganic contaminants (As: 0.5; Cd: 0.5; Cr: 0.1; Cu: 30; Ni: 3; Pb: 0.4; Sb: 1 and Zn: 25 mg L-1). Determination of arsenic species has shown inorganic As(V) as predominant in most samples. Sodium concentrations, nowadays a major public health concern, were also measured, showing an average of 149 mg L-1. Analytical results for this element were much higher than label concentrations, showing the need for better quality control.

Matéria orgânica em quatro tipos de solos brasileiros: composição química e sorção de atrazina

Soil organic matter is the main sorptive soil compartment for atrazine in soils, followed in a minor scale by the inorganic fraction. In this study, the soil organic matter quality and atrazine sorption were investigated in four different soil types. The pedogenic environment affected the humification and therefore the chemical composition of the organic matter. The organic matter contribution to atrazine sorption was larger (60-83%) than that of the inorganic fraction. The organic matter capacity in retaining the herbicide was favoured by a higher decomposition degree and a smaller carboxylic substitution of the aliphatic chains.

Homogeneous Assays for Simplified Screening of HLA-conferred Genetic Disease Risk

The increasing incidence of type 1 diabetes has led researchers on a quest to find the reason behind this phenomenon. The rate of increase is too great to be caused simply by changes in the genetic component, and many environmental factors are under investigation for their possible contribution. These studies require, however, the participation of those individuals most likely to develop the disease, and the approach chosen by many is to screen vast populations to find persons with increased genetic risk factors. The participating individuals are then followed for signs of disease development, and their exposure to suspected environmental factors is studied. The main purpose of this study was to find a suitable tool for easy and inexpensive screening of certain genetic risk markers for type 1 diabetes. The method should be applicable to using whole blood dried on sample collection cards as sample material, since the shipping and storage of samples in this format is preferred. However, the screening of vast sample libraries of extracted genomic DNA should also be possible, if such a need should arise, for example, when studying the effect of newly discovered genetic risk markers. The method developed in this study is based on homogeneous assay chemistry and an asymmetrical polymerase chain reaction (PCR). The generated singlestranded PCR product is probed by lanthanide-labelled, LNA (locked nucleic acid)-spiked, short oligonucleotides with exact complementary sequences. In the case of a perfect match, the probe is hybridised to the product. However, if even a single nucleotide difference occurs, the probe is bound instead of the PCR product to a complementary quencher-oligonucleotide labelled with a dabcyl-moiety, causing the signal of the lanthanide label to be quenched. The method was applied to the screening of the well-known type 1 diabetes risk alleles of the HLA-DQB1 gene. The method was shown to be suitable as an initial screening step including thousands of samples in the scheme used in the TEDDY (The Environmental Determinants of Diabetes in the Young) study to identify those individuals at increased genetic risk. The method was further developed into dry-reagent form to allow an even simpler approach to screening. The reagents needed in the assay were in dry format in the reaction vessel, and performing the assay required only the addition of the sample and, if necessary, water to rehydrate the reagents. This allows the assay to be successfully executed even by a person with minimal laboratory experience.

O sistema globalmente harmonizado de classificação e rotulagem de produtos químicos (GHS): uma introdução para sua aplicação em laboratórios de ensino e pesquisa acadêmica

The present article provides an overview of the Globally Harmonised System of Classification, Labelling and Packaging of Chemicals (GHS) and its implementation in Brazil. Although Classification and Packaging is beyond the scope of the responsibility of academic chemists, labelling of chemicals used in academic laboratories will be required by the competent authorities to ensure the safety of students and staff. Therefore, academic teachers and researchers responsible should be familiarised with the GHS principles outlined here and at least be able to label, by applying these principles, mixtures of substances previously classified by the competent authorities.