Algorithmic Techniques in Gene Expression Processing. From Imputation to Visualization

The amount of biological data has grown exponentially in recent decades. Modern biotechnologies, such as microarrays and next-generation sequencing, are capable to produce massive amounts of biomedical data in a single experiment. As the amount of the data is rapidly growing there is an urgent need for reliable computational methods for analyzing and visualizing it. This thesis addresses this need by studying how to efficiently and reliably analyze and visualize high-dimensional data, especially that obtained from gene expression microarray experiments. First, we will study the ways to improve the quality of microarray data by replacing (imputing) the missing data entries with the estimated values for these entries. Missing value imputation is a method which is commonly used to make the original incomplete data complete, thus making it easier to be analyzed with statistical and computational methods. Our novel approach was to use curated external biological information as a guide for the missing value imputation. Secondly, we studied the effect of missing value imputation on the downstream data analysis methods like clustering. We compared multiple recent imputation algorithms against 8 publicly available microarray data sets. It was observed that the missing value imputation indeed is a rational way to improve the quality of biological data. The research revealed differences between the clustering results obtained with different imputation methods. On most data sets, the simple and fast k-NN imputation was good enough, but there were also needs for more advanced imputation methods, such as Bayesian Principal Component Algorithm (BPCA). Finally, we studied the visualization of biological network data. Biological interaction networks are examples of the outcome of multiple biological experiments such as using the gene microarray techniques. Such networks are typically very large and highly connected, thus there is a need for fast algorithms for producing visually pleasant layouts. A computationally efficient way to produce layouts of large biological interaction networks was developed. The algorithm uses multilevel optimization within the regular force directed graph layout algorithm.

Vibrations in Rotating Machinery Arising from Minor Imperfections in Component Geometries

Vibrations in machines can cause noise, decrease the performance, or even damage the machine. Vibrations appear if there is a source of vibration that excites the system. In the worst case scenario, the excitation frequency coincides with the natural frequency of the machine causing resonance. Rotating machines are a machine type, where the excitation arises from the machine itself. The excitation originates from the mass imbalance in the rotating shaft, which always exists in machines that are manufactured using conventional methods. The excitation has a frequency that is dependent on the rotational speed of the machine. The rotating machines in industrial use are usually designed to rotate at a constant rotational speed, the case where the resonances can be easily avoided. However, the machines that have a varying operational speed are more problematic due to a wider range of frequencies that have to be avoided. Vibrations, which frequencies equal to rotational speed frequency of the machine are widely studied and considered in the typical machine design process. This study concentrates on vibrations, which arise from the excitations having frequencies that are multiples of the rotational speed frequency. These vibrations take place when there are two or more excitation components in a revolution of a rotating shaft. The dissertation introduces four studies where three kinds of machines are experiencing vibrations caused by different excitations. The first studied case is a directly driven permanent magnet generator used in a wind power plant. The electromagnetic properties of the generator cause harmonic excitations in the system. The dynamic responses of the generator are studied using the multibody dynamics formulation. In another study, the finite element method is used to study the vibrations of a magnetic gear due to excitations, which frequencies equal to the rotational speed frequency. The objective is to study the effects of manufacturing and assembling inaccuracies. Particularly, the eccentricity of the rotating part with respect to non-rotating part is studied since the eccentric operation causes a force component in the direction of the shortest air gap. The third machine type is a tube roll of a paper machine, which is studied while the tube roll is supported using two different structures. These cases are studied using different formulations. In the first case, the tube roll is supported by spherical roller bearings, which have some wavinesses on the rolling surfaces. Wavinesses cause excitations to the tube roll, which starts to resonate at the frequency that is a half of the first natural frequency. The frequency is in the range where the machine normally operates. The tube roll is modeled using the finite element method and the bearings are modeled as nonlinear forces between the tube roll and the pedestals. In the second case studied, the tube roll is supported by freely rotating discs, which wavinesses are also measured. The above described phenomenon is captured as well in this case, but the simulation methodology is based on the flexible multibody dynamics formulation. The simulation models that are used in both of the last two cases studied are verified by measuring the actual devices and comparing the simulated and measured results. The results show good agreement.

Enhancement of ultrafiltration process by pretreatment in recovery of hemicelluloses from wood extracts

Hemicelluloses are potential raw material for several items produced in future wood-based biorefineries. One possible method for recovering hemicelluloses from wood extracts is ultrafiltration (UF). However, low filtration capacities and severe fouling restrict the use of tight UF membranes in the treatment of wood extracts. The lack of suitable commercial membranes creates a need for pretreatment which would decrease fouling and increase the filtration capacity. This thesis focuses on the evaluation of the possibility to improve the filtration capacity and decrease fouling with the pretreatment of wood extracts. Methods which remove harmful compounds and methods which degrade them are studied, as well as combinations of the methods. The tested pretreatments have an influence on both the concentration of different compounds and the molecular mass distribution of the compounds in the extract. This study revealed that in addition to which kind of compounds were removed, also the change in molecular size distribution affected the filtration capacity significantly. It was shown that the most harmful compounds for the filtration capacity of the hydrophobic 5 kDa membrane were the ones capable of permeating the membrane and fouling also the inner membrane structure. Naturally, the size of the most harmful compounds depends on the used UF membrane and is thus case-specific. However, in the choice of the pretreatment method, the focus should be on the removal of harmful compound sizes rather than merely on the total amount of removed foulants. The results proved that filtration capacity can be increased with both adsorptive and oxidative pretreatments even by hundreds of per cents. For instance, the use of XAD7 and XAD16 adsorbents increased the average flux in the UF of a birch extract from nearly zero to 107 kg/(m2h) and 175 kg/(m2h), respectively. In the treatment of a spruce extract, oxidation by pulsed corona discharge (PCD) increased the flux in UF from 46 kg/(m2h) to 158 kg/(m2h). Moreover, when a birch extract batch was treated with laccase enzyme, the flux in UF increased from 15 kg/(m2h) to 36 kg/(m2h). However, fouling was decreased only by adsorptive pretreatment while oxidative methods had a negligible or even negative impact on it. This demonstrates that filtration capacity and fouling are affected by different compounds and mechanisms. The results of this thesis show that filtration capacity can be improved and fouling decreased through appropriate pretreatment. However, the choice of the best possible pretreatment is case-specific and depends on the wood extract and the membrane used. Finding the best option requires information on the extract content and membrane characteristics as well as on the filtration performance of the membrane in the prevailing conditions and a multivariate approach. On the basis of this study, it can be roughly concluded that adsorptive pretreatment improves the filtration capacity and decreases fouling rather reliably, but it may lead to significant hemicellulose losses. Oxidation reduces the loss of valuable hemicelluloses and could improve the filtration capacity, but fouling challenges may remain. Combining oxidation with adsorptive pretreatment was not a solution for avoiding hemicellulose losses in the tested cases.

Corporate social responsibility in online retail – Perceptions from online discussions on Amazon

Online retail has experienced substantial growth due to the increased utilization of digital technologies. The growth in the industry has created numerous multinational corporations, which serve customers all around the world. As a result, the online retailers’ efforts in sustainability have been more often in publicity. Corporate Social Responsibility (CSR), the efforts going beyond the company’s interest in order to bring social good, have been discussed in the academic community for decades. While the various effects of CSR in retail and business generally have been realized, the research aims to make a contribution by examining the concept from the viewpoint of online retail by focusing on interaction at online discussion forums. The main research question asks: What is the role of online retailer CSR in the online discussion forum context? In order to answer this question, additional three sub-research questions are suggested: 1) What elements of CSR are relevant in the online retailer context? 2) How are discussion forums used in communication regarding CSR and online retail by different discusser types? 3) What kind of company perceptions and consumer behavior appear in online discussions of Amazon’s CSR? The study is qualitative of nature, using the content analysis research method to examine posts found from four different public online discussion forums. A qualitative, interpretative analysis was used, although quantitative measures regarding discussion posts were also presented. The analysis contains a coding process, where attributes are attached to discussion posts. As a result of the process, specific patterns can be indicated in order to categorize the posts. As the research uses Amazon as the case company, the case study method is also adopted. In the online retail context, ethics regarding the online retailer’s tax and labor policies are found as most relevant elements. Customers, using discussion forums to share their experiences and opinions regarding online retailer CSR efforts, are realized as the most prominent stakeholder. The research also highlights the role of online retailer employees providing details on internal business practices. However, the not realize the possibility to participate in discussions. Most discussion posts refer either to a positive, negative or mixed reaction towards CSR efforts, it is suggested that the concept is relevant also in online retail. Although online retailer CSR efforts are somewhat linked to company perceptions, the concept has a minimal role in consumer behavior. In online retail, CSR efforts can reduce the risk of being involved in discussions related to controversies. In addition, online retailer CSR efforts should communicate how the retailer contributes locally.

Corporate social responsibility in online retail – Perceptions from online discussions on Amazon

Online retail has experienced substantial growth due to the increased utilization of digital technologies. The growth in the industry has created numerous multinational corporations, which serve customers all around the world. As a result, the online retailers’ efforts in sustainability have been more often in publicity. Corporate Social Responsibility (CSR), the efforts going beyond the company’s interest in order to bring social good, have been discussed in the academic community for decades. While the various effects of CSR in retail and business generally have been realized, the research aims to make a contribution by examining the concept from the viewpoint of online retail by focusing on interaction at online discussion forums. The main research question asks: What is the role of online retailer CSR in the online discussion forum context? In order to answer this question, additional three sub-research questions are suggested: 1) What elements of CSR are relevant in the online retailer context? 2) How are discussion forums used in communication regarding CSR and online retail by different discusser types? 3) What kind of company perceptions and consumer behavior appear in online discussions of Amazon’s CSR? The study is qualitative of nature, using the content analysis research method to examine posts found from four different public online discussion forums. A qualitative, interpretative analysis was used, although quantitative measures regarding discussion posts were also presented. The analysis contains a coding process, where attributes are attached to discussion posts. As a result of the process, specific patterns can be indicated in order to categorize the posts. As the research uses Amazon as the case company, the case study method is also adopted. In the online retail context, ethics regarding the online retailer’s tax and labor policies are found as most relevant elements. Customers, using discussion forums to share their experiences and opinions regarding online retailer CSR efforts, are realized as the most prominent stakeholder. The research also highlights the role of online retailer employees providing details on internal business practices. However, the not realize the possibility to participate in discussions. Most discussion posts refer either to a positive, negative or mixed reaction towards CSR efforts, it is suggested that the concept is relevant also in online retail. Although online retailer CSR efforts are somewhat linked to company perceptions, the concept has a minimal role in consumer behavior. In online retail, CSR efforts can reduce the risk of being involved in discussions related to controversies. In addition, online retailer CSR efforts should communicate how the retailer contributes locally.