Studying the internal characteristics of a network using measurements obtained from endhosts is known as network tomography. The foremost challenge in measurement-based approaches is the large size of a network, where only a subset of measurements can be obtained because of the inaccessibility of the entire network. As the network becomes larger, a question arises as to how rapidly the monitoring resources (number of measurements or number of samples) must grow to obtain a desired monitoring accuracy. Our work studies the scalability of the measurements with respect to the size of the network. We investigate the issues of scalability and performance evaluation in IP networks, specifically focusing on fault and congestion diagnosis. We formulate network monitoring as a machine learning problem using probabilistic graphical models that infer network states using path-based measurements. We consider the theoretical and practical management resources needed to reliably diagnose congested/faulty network elements and provide fundamental limits on the relationships between the number of probe packets, the size of the network, and the ability to accurately diagnose such network elements. We derive lower bounds on the average number of probes per edge using the variational inference technique proposed in the context of graphical models under noisy probe measurements, and then propose an entropy lower (EL) bound by drawing similarities between the coding problem over a binary symmetric channel and the diagnosis problem. Our investigation is supported by simulation results. For the congestion diagnosis case, we propose a solution based on decoding linear error control codes on a binary symmetric channel for various probing experiments. To identify the congested nodes, we construct a graphical model, and infer congestion using the belief propagation algorithm. In the second part of the work, we focus on the development of methods to automatically analyze the information contained in electron tomograms, which is a major challenge since tomograms are extremely noisy. Advances in automated data acquisition in electron tomography have led to an explosion in the amount of data that can be obtained about the spatial architecture of a variety of biologically and medically relevant objects with sizes in the range of 10-1000 nm A fundamental step in the statistical inference of large amounts of data is to segment relevant 3D features in cellular tomograms. Procedures for segmentation must work robustly and rapidly in spite of the low signal-to-noise ratios inherent in biological electron microscopy. This work evaluates various denoising techniques and then extracts relevant features of biological interest in tomograms of HIV-1 in infected human macrophages and Bdellovibrio bacterial tomograms recorded at room and cryogenic temperatures. Our approach represents an important step in automating the efficient extraction of useful information from large datasets in biological tomography and in speeding up the process of reducing gigabyte-sized tomograms to relevant byte-sized data. Next, we investigate automatic techniques for segmentation and quantitative analysis of mitochondria in MNT-1 cells imaged using ion-abrasion scanning electron microscope, and tomograms of Liposomal Doxorubicin formulations (Doxil), an anticancer nanodrug, imaged at cryogenic temperatures. A machine learning approach is formulated that exploits texture features, and joint image block-wise classification and segmentation is performed by histogram matching using a nearest neighbor classifier and chi-squared statistic as a distance measure.

The aim of the thesis is to verify synergy of genetic programming and neural networks. Solution is provided by set of experiments with implemented library built upon benchmark tasks. I've done experiments with directly and also indirectly encoded neural netwrok. I focused on finding robust solutions and the best calculation of configurations, overfitting detection and advanced stimulations of solution with fitness function. Generally better solutions were found using lower values of parameters n_c and n_r. These solutions tended less to be overfitted. I was able to evolve neurocontroller eliminating oscilations in pole balancing problem. In cancer detection problem, precision of provided solution was over 98%, which overcame compared techniques. I succeeded also in designing of maze model, where agent was able to perform multistep tasks.

Under senare år har målinriktad terapi varit ett växande område inom cancerterapi som en mer målinriktad behandling än kemoterapi. Dessa behandlingar baseras främst på antikroppsbaserade läkemedel som är ganska stora och komplexa, vilket ökar den totala kostnaden för behandlingen. Därför måste man hitta en alternativ metod för både upptäckt och behandling för att hjälpa patienterna. Små affinitetsdomäner har skapats med målet att förbättra vävnadspenetrationen och samtidigt upprätthålla en hög grad av målspecificitet, vilket leder till färre biverkningar. Ett av exemplen på detta är Albumin Binding Domain-Derived Affinity Protein (ADAPT). Det har baserats på en av de albuminbindande domänerna (ABD) i streptokockproteinet G, med en storlek på 6,5 kDa. Nyligen modifierades ADAPT ytterligare för att samtidigt binda albumin och ett annat relevant målprotein av intresse, vilket tyder på en längre halveringstid i patientserum och möjliggör utveckling av nyare och terapeutiska läkemedel. I detta projekt presenteras den fjärde generationen av ADAPT-biblioteket som utformats för att ha förbättrad stabilitet. Selektioner med fagdisplay utfördes mot tre målproteiner: carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5), en biomarkör för kolorektalcancer, epithelial cell adhesion molecule (EpCAM), en markör för flera gastrointestinala karcinom och trophoblast cell-surface antigen 2 (Trop2) som är överuttryckt i trippel-negativ bröstcancer. Resultaten visar bindning till CEACAM5, EpCAM och Trop2, vilket har visats med monoklonal fag-ELISA. Bindningsaffiniteten, sekundärstrukturen hos de utvalda bindarna och den bispecifika bindningen till serumalbumin återstår att utvärdera ytterligare. Projektet visar således att de ADAPTs som valts ut mot målen CEACAM5, EpCAM och Trop2 har en enorm potential för framtida kliniska tillämpningar som syftar till utveckling av diagnostik och terapi för dessa cancerbiomarkörer. / In recent years, targeted therapy has been a growing field of cancer therapy as a more targeted treatment than chemotherapy. These treatments are primarily based on antibody-based pharmaceuticals which are quite large and complex, increasing the overall cost of the treatment. Thus, an alternative method of both detection and treatment needs to be found to aid patients. Small affinity domains have been created with the goal of enhancing tissue penetration while maintaining a high level of target specificity, leading to fewer side effects. One of the examples for these is the Albumin Binding Domain-Derived Affinity Protein (ADAPT). It has been based on one of the albumin binding domains (ABD) of the streptococcal protein G, with a size of 6.5 kDa. Recently, the ADAPTs were further modified to simultaneously bind albumin and another pertinent target protein of interest, suggesting a longer half-life in patient serum, and enabling the development of newer therapeutics. This project presents the 4th generation of the ADAPT library designed to have improved stability. Phage display selections were performed against three target proteins: carcinoembryonic antigen- related cell adhesion molecule 5 (CEACAM5), a biomarker for colorectal cancer, epithelial cell adhesion molecule (EpCAM), a marker for several gastrointestinal carcinomas and trophoblast cell-surface antigen 2 (Trop2) which is overexpressed in triple-negative breast cancer. The results demonstrate binding towards CEACAM5, EpCAM and Trop2, which has been shown by monoclonal phage ELISA. The binding affinity, secondary structure of the selected binders and bispecific binding towards serum albumin remain to be further assessed. The project thus reveals that the ADAPTs selected against the targets CEACAM5, EpCAM and Trop2 present a massive potential for future clinical applications aimed towards development of diagnostics and therapeutics for these cancer biomarkers.