L’objectif principal de ce projet de recherche doctorale est la mise au point d’approches méthodologiques fiables et reproductibles permettant la caractérisation génétique de cellules rares circulantes isolées par la méthode de filtration ISET® (Rarecells®, France). La première application développée consiste en la détection des mutations du gène suppresseur de tumeur VHL (Von Hippel Lindau) dans les cellules rares circulantes (CRC) uniques isolées du sang de 30 patients atteints de carcinome rénal à cellules claires (CRCC), et réalisée comparativement à l’analyse cytopathologique. L’étude génétique a également été conduite en parallèle dans les 30 tissus tumoraux correspondants. Les résultats ont mis en lumière une potentielle complémentarité de l’approche de génétique moléculaire sur cellules uniques avec l’analyse cytomorphologique de référence et suggèrent que combiner ces approches permettrait d’obtenir une plus grande sensibilité de détection des cellules cancéreuses circulantes chez les patients atteints de CRCC. Une deuxième application a consisté en le développement d’une approche innovante pour le diagnostic prénatal non-invasif des maladies génétiques récessives par analyse de cellules trophoblastiques rares collectées au niveau du col de l’utérus. Enfin, des développements supplémentaires ont permis d’optimiser les analyses de séquençage à haut débit et de les appliquer à des CRC individuelles isolées par ISET®. Cette nouvelle approche, associée à l’isolement de CRC non fixées, est en mesure de fournir des données génétiques élargies à l’exome entier pour des applications à la fois en oncologie prédictive et en diagnostic prénatal non invasif. / The aim of this doctoral research project is the development of reliable and reproducible methodological approaches enabling the genetic characterization of circulating rare cells (CRC) isolated by ISET® filtration (Rarecells®, France). The first application developed consists in detecting mutations of the VHL (Von Hippel Lindau) tumor suppressor gene in single CRC isolated from the blood of 30 patients patients with clear cell renal cell carcinoma (ccRCC), assessed according to the results obtained by cytopathological analysis. In parallel, genetic analysis of VHL mutations was conducted in the corresponding tumor tissues. Results revealed a potential complementarity of the molecular genetic approach targeted to single cells with the reference method of cytopathological analysis and suggested that combining both strategies could improve the sensitivity of circulating cancer cells’ detection in patients with ccRCC. A second application consisted in the development of an innovative approach for non-invasive prenatal diagnosis of recessive genetic diseases by analysis of rare trophoblastic cells collected from the cervix. Finally, further developments allowed to optimize high-throughput sequencing analyses and to apply them to single CRC isolated by ISET®. This approach, combined with the isolation of living CRC, enabled us to obtain broader genetic data from the whole exome and should foster innovative applications to both predictive oncology and non-invasive prenatal diagnosis.

This assignment is based on the tolerance problems that occur with the radial dimensions or RR-dimensions for the windings. The purpose is to identify the various parameters that affect said RR-dimensions and where these parameters originate from to be able to assure quality of the winding process. The RR-dimension has a tolerance of + 2,2 and - 1,1 mm, but it is only the positive tolerance that cause problems. The case study is based on a mixture of quantitative and qualitative methods. Quantitative data collections of dimensions for leaders and how the RR-measurement are affected by them, analysis of tolerances and how well they are adapted for their purpose. Qualitative observations, experiments and semi-structured interviews have also been conducted to gain a deeper understanding of the problem. The results show that tolerance problems originate from several different factors. The first factor is the debatable tolerance setting on some conductors, mainly the thin CTC conductors that do not have a high radial dimension. These conductors have tolerances that do not fill a purpose because it does not ensure the quality for the winding. The conductors need a tighter tolerance spectrum for the positive direction to be fully functional. The air gap that can occur between the conductors is also a cause of error. This stems from the fact that it is not possible to achieve a sufficiently high force when tensioning the disc. This can also occur from the uneven surface of the CTC-conductor that can create gaps. Another cause of error that affects the dimension is the number of turns for a disc. When a disc is wound with a high number of turns it will lead tothat the influencing factors have a higher amount of turns to effect on. This means that a small increase of the conductor's nominal dimension or small air gaps has a large impact on the overall RR-dimension. / För att kvalitetssäkra transformator- och reaktorlindningarna hos företaget Hitachi ABB skade utsatta toleranserna för varje lindning uppfyllas för att kunna fortsätta vidare i produktionslinan. Arbetet är grundat utifrån de förekommande variationerna hos lindningarnas mått i radiell riktning (RR-mått. Arbetets syfte är att undersöka Hitachi ABB:s lindningsproduktion för att finna de faktorer som ger upphov till måttvariationer hos RRmåttet. De identifierade faktorerna ska användas för att kvalitetssäkra företagets lindningsprocess. Färdigställt arbete ska användas som grund för företagets fortsatta arbetekring kvalitetssäkring. RR-måttet har en tolerans på + 2,2 och – 1,1 mm, däremot är detendast de variationer i positivt led där problem uppstår. Vid ett för högt mått genomförs enutredning för att se över om lindningen kan brukas ändå eller utifall de för höga måttet i radiell riktning påverkar prestandan samt montering. Arbetet är en fallstudie utfört med en blandning av kvantitativa och kvalitativa metoder. Denkvantitativa datainsamlingens syfte är att undersöka ledares mått i lindningar för att se hur RR-måttet påverkas måttsättning och varvantal. Den kvantitativa datainsamlingen genomför också analyser av toleranser och hur väl anpassade de är för sitt syfte. Kvalitativa observationer, experiment och halvstrukturerade interjuver har genomförts för att skapa en djupare förståelse kring problemet.Resultatet visar att de variationer som uppstår hos RR-måttet påverkas av flera olika faktorer.Variationer hos RR-måttet uppstår främst för lindningar där en CTC-ledare används. En påverkande faktor är toleransvidden för ledarna, främst för de tunna CTC-ledarna. Ledarna lindas runt varandra där ledarens mått staplas på varandra för att bilda ett RR-mått. För de tunna CTC-ledarna bör en mindre toleransvidd användas, mer specifikt en lägre positiv tolerans likt de platta ledarna för att kunna kvalitetssäkra lindningen. En annan påverkandefaktor är de luftspalter som uppstår mellan ledarna. Luftspalter härstammar ifrån att det integår att uppnå tillräckligt hög kraft vid spänning av en skiva alternativt från CTC-ledarens knöliga yta. Utöver dessa två faktorer har antalet lindade varv för en skiva en påverkan förtoleransstapling. Fler varv resulterar i att små måttavvikelser för ledare påverkar det totala måttet, vilket innebär att de luftspalter som kan uppstå blir fler när antalet varv ökar.

No description available.

This master thesis deals with the implementation of various types of recurrent neural networks via programming language lua using torch library. It focuses on finding optimal strategy for training recurrent neural networks and also tries to minimize the duration of the training. Furthermore various types of regularization techniques are investigated and implemented into the recurrent neural network architecture. Implemented recurrent neural networks are compared on the speech recognition task using AMI dataset, where they model the acustic information. Their performance is also compared to standard feedforward neural network. Best results are achieved using BLSTM architecture. The recurrent neural network are also trained via CTC objective function on the TIMIT dataset. Best result is again achieved using BLSTM architecture.

Colorectal cancer (CRC) is the third most common cancer worldwide; it is responsible for nearly 10% of all newly diagnosed cancers and is the second most cause of cancer related death in Europe. Biomarkers for therapy guidance, targeted therapy and survival prognosis are still limited. As CRC is a heterogeneous disease, different parts of the tumor might have varying molecular characteristics which may change during therapy or disease progression. Through solid biopsies and screenings, these local or temporal differences are impossible to monitor. To facilitate detection of these possible temporal changes, a regularly and non-invasively accessible biomarker is required for disease monitoring. Circulating tumor cells (CTCs) might represent such a biomarker as they have been shown to be fluid surrogates of the solid tumor. EpCAM positive CTCs have shown to be prognostic in CRC for survival, but their full potential has not yet been evaluated further. By using the High Definition Single Cell Analysis (HD-SCA) workflow, we were able to analyze the entire spectrum of CTCs and categorize them as the regular CTCs (HD-CTC), CTCs with a smaller nuclear area (CTC-Small), CTCs with low expression of epithelial marker cytokeratin (CTC-LowCK) and CTCs undergoing apoptosis and therefore releasing cell free DNA...

Objective: The aim of this study was to develop an immunomagnetic/real-time reverse transcriptase polymerase chain reaction (RT-PCR) assay and assess its clinical value for the molecular detection of circulating tumor cells (CTCs) in peripheral blood of pancreatic cancer patients. Methods: The presence of CTCs was evaluated in 34 pancreatic cancer patients before systemic therapy and in 40 healthy controls, through immunomagnetic enrichment, using the antibodies BM7 and VU1D9 [targeting mucin 1 and epithelial cell adhesion molecule (EpCAM), respectively], followed by real-time RT-PCR analysis of the genes KRT19, MUC1, EPCAM, CEACAM5 and BIRC5. Results: The developed assay showed high specificity, as none of the healthy controls were found to be positive for the multimarker gene panel. CTCs were detected in 47.1% of the pancreatic cancer patients before the beginning of systemic treatment. Shorter median progression-free survival (PFS) was observed for patients who had at least one detectable tumor-associated transcript, compared with patients who were CTC negative. Median PFS time was 66.0 days [95% confidence interval (CI) 44.8–87.2] for patients with baseline CTC positivity and 138.0 days (95% CI 124.1–151.9) for CTC-negative patients (p = 0.01, log-rank test). Conclusion: Our results suggest that in addition to the current prognostic methods, CTC analysis represents a potential complementary tool for prediction of outcome in pancreatic cancer patients. / Dieser Beitrag ist mit Zustimmung des Rechteinhabers aufgrund einer (DFG-geförderten) Allianz- bzw. Nationallizenz frei zugänglich.

In Automatic Speech Recognition, there is an extensive need for time-aligned data. Manual speech segmentation has been shown to be more laborious than manual transcription, especially when dealing with tens of hours of speech. Forced alignment is a technique for matching a signal with its orthographic transcription with respect to the duration of linguistic units. Most forced aligners, however, are language-dependent and trained on English data, whereas under-resourced languages lack the resources to develop an acoustic model required for an aligner, as well as manually aligned data. An alternative solution to the training of new models can be cross-language forced alignment, in which an aligner trained on one language is used for aligning data in another language.  This thesis aimed to evaluate state-of-the-art forced alignment algorithms available for Swedish and test whether a Swedish model could be applied for aligning Norwegian. Three approaches for forced aligners were employed: (1) one forced aligner based on Dynamic Time Warping and text-to-speech synthesis Aeneas, (2) two forced aligners based on Hidden Markov Models, namely the Munich AUtomatic Segmentation System (WebMAUS) and the Montreal Forced Aligner (MFA) and (3) Connectionist Temporal Classification (CTC) segmentation algorithm with two pre-trained and fine-tuned Wav2Vec2 Swedish models. First, small speech test sets for Norwegian and Swedish, covering different types of spontaneousness in the speech, were created and manually aligned to create gold-standard alignments. Second, the performance of the Swedish dataset was evaluated with respect to the gold standard. Finally, it was tested whether Swedish forced aligners could be applied for aligning Norwegian data. The performance of the aligners was assessed by measuring the difference between the boundaries set in the gold standard from that of the comparison alignment. The accuracy was estimated by calculating the proportion of alignments below a particular threshold proposed in the literature. It was found that the performance of the CTC segmentation algorithm with Wav2Vec2 (VoxRex) was superior to other forced alignment systems. The differences between the alignments of two Wav2Vec2 models suggest that the training data may have a larger influence on the alignments, than the architecture of the algorithm. In lower thresholds, the traditional HMM approach outperformed the deep learning models. Finally, findings from the thesis have demonstrated promising results for cross-language forced alignment using Swedish models to align related languages, such as Norwegian.

Colorectal cancer (CRC) is the third most common cancer worldwide; it is responsible for nearly 10% of all newly diagnosed cancers and is the second most cause of cancer related death in Europe. Biomarkers for therapy guidance, targeted therapy and survival prognosis are still limited. As CRC is a heterogeneous disease, different parts of the tumor might have varying molecular characteristics which may change during therapy or disease progression. Through solid biopsies and screenings, these local or temporal differences are impossible to monitor. To facilitate detection of these possible temporal changes, a regularly and non-invasively accessible biomarker is required for disease monitoring. Circulating tumor cells (CTCs) might represent such a biomarker as they have been shown to be fluid surrogates of the solid tumor. EpCAM positive CTCs have shown to be prognostic in CRC for survival, but their full potential has not yet been evaluated further. By using the High Definition Single Cell Analysis (HD-SCA) workflow, we were able to analyze the entire spectrum of CTCs and categorize them as the regular CTCs (HD-CTC), CTCs with a smaller nuclear area (CTC-Small), CTCs with low expression of epithelial marker cytokeratin (CTC-LowCK) and CTCs undergoing apoptosis and therefore releasing cell free DNA...

No description available.