Statistical learning theory for geospatial data. Case study: Aral sea

In recent years there has been an explosive growth in the development of adaptive and data driven methods. One of the efficient and data-driven approaches is based on statistical learning theory (Vapnik 1998). The theory is based on Structural Risk Minimisation (SRM) principle and has a solid statistical background. When applying SRM we are trying not only to reduce training error ? to fit the available data with a model, but also to reduce the complexity of the model and to reduce generalisation error. Many nonlinear learning procedures recently developed in neural networks and statistics can be understood and interpreted in terms of the structural risk minimisation inductive principle. A recent methodology based on SRM is called Support Vector Machines (SVM). At present SLT is still under intensive development and SVM find new areas of application (www.kernel-machines.org). SVM develop robust and non linear data models with excellent generalisation abilities that is very important both for monitoring and forecasting. SVM are extremely good when input space is high dimensional and training data set i not big enough to develop corresponding nonlinear model. Moreover, SVM use only support vectors to derive decision boundaries. It opens a way to sampling optimization, estimation of noise in data, quantification of data redundancy etc. Presentation of SVM for spatially distributed data is given in (Kanevski and Maignan 2004).

Structure-function characterization and optimization of a plant-derived antibacterial peptide.

Crushed seeds of the Moringa oleifera tree have been used traditionally as natural flocculants to clarify drinking water. We previously showed that one of the seed peptides mediates both the sedimentation of suspended particles such as bacterial cells and a direct bactericidal activity, raising the possibility that the two activities might be related. In this study, the conformational modeling of the peptide was coupled to a functional analysis of synthetic derivatives. This indicated that partly overlapping structural determinants mediate the sedimentation and antibacterial activities. Sedimentation requires a positively charged, glutamine-rich portion of the peptide that aggregates bacterial cells. The bactericidal activity was localized to a sequence prone to form a helix-loop-helix structural motif. Amino acid substitution showed that the bactericidal activity requires hydrophobic proline residues within the protruding loop. Vital dye staining indicated that treatment with peptides containing this motif results in bacterial membrane damage. Assembly of multiple copies of this structural motif into a branched peptide enhanced antibacterial activity, since low concentrations effectively kill bacteria such as Pseudomonas aeruginosa and Streptococcus pyogenes without displaying a toxic effect on human red blood cells. This study thus identifies a synthetic peptide with potent antibacterial activity against specific human pathogens. It also suggests partly distinct molecular mechanisms for each activity. Sedimentation may result from coupled flocculation and coagulation effects, while the bactericidal activity would require bacterial membrane destabilization by a hydrophobic loop.

Uncertainty Quantification with Support Vector Regression Prediction Models

Uncertainty quantification of petroleum reservoir models is one of the present challenges, which is usually approached with a wide range of geostatistical tools linked with statistical optimisation or/and inference algorithms. The paper considers a data driven approach in modelling uncertainty in spatial predictions. Proposed semi-supervised Support Vector Regression (SVR) model has demonstrated its capability to represent realistic features and describe stochastic variability and non-uniqueness of spatial properties. It is able to capture and preserve key spatial dependencies such as connectivity, which is often difficult to achieve with two-point geostatistical models. Semi-supervised SVR is designed to integrate various kinds of conditioning data and learn dependences from them. A stochastic semi-supervised SVR model is integrated into a Bayesian framework to quantify uncertainty with multiple models fitted to dynamic observations. The developed approach is illustrated with a reservoir case study. The resulting probabilistic production forecasts are described by uncertainty envelopes.

A cell therapy approach for erythropoietin delivery using encapsulated human primary fibroblasts

Summary Cell therapy has emerged as a strategy for the treatment of various human diseases. Cells can be transplanted considering their morphological and functional properties to restore a tissue damage, as represented by blood transfusion, bone marrow or pancreatic islet cells transplantation. With the advent of the gene therapy, cells also were used as biological supports for the production of therapeutic molecules that can act either locally or at distance. This strategy represents the basis of ex vivo gene therapy characterized by the removal of cells from an organism, their genetic modification and their implantation into the same or another individual in a physiologically suitable location. The tissue or biological function damage dictates the type of cells chosen for implantation and the required function of the implanted cells. The general aim of this work was to develop an ex vivo gene therapy approach for the secretion of erythropoietin (Epo) in patients suffering from Epo-responsive anemia, thus extending to humans, studies previously performed with mouse cells transplanted in mice and rats. Considering the potential clinical application, allogeneic primary human cells were chosen for practical and safety reasons. In contrast to autologous cells, the use of allogeneic cells allows to characterize a cell lineage that can be further transplanted in many individuals. Furthermore allogeneic cells avoid the potential risk of zoonosis encountered with xenogeneic cells. Accordingly, the immune reaction against this allogeneic source was prevented by cell macro- encapsulation that prevents cell-to-cell contact with the host immune system and allows to easy retrieve the implanted device. The first step consisted in testing the survival of various human primary cells that were encapsulated and implanted for one month in the subcutaneous tissue of immunocompetent and naturally or therapeutically immunodepressed mice, assuming that xenogeneic applications constitute a stringent and representative screening before human transplantation. A fibroblast lineage from the foreskin of a young donor, DARC 3.1 cells, showed the highest mean survival score. We have then performed studies to optimize the manufacturing procedures of the encapsulation device for successful engraftment. The development of calcifications on the polyvinyl alcohol (PVA) matrix serving as a scaffold for enclosed cells into the hollow fiber devices was reported after one month in vivo. Various parameters, including matrix rinsing solutions, batches of PVA and cell lineages were assessed for their respective role in the development of the phenomenon. We observed that the calcifications could be totally prevented by using ultra-pure sterile water instead of phosphate buffer saline solution in the rinsing procedure of the PVA matrix. Moreover, a higher lactate dehydrogenase activity of the cells was found to decrease calcium depositions due to more acidic microenvironment, inhibiting the calcium precipitation. After the selection of the appropriate cell lineage and the optimization of encapsulation conditions, a retroviral-based approach was applied to DARC 3.1 fibroblasts for the transduction of the human Epo cDNA. Various modifications of the retroviral vector and the infection conditions were performed to obtain clinically relevant levels of human Epo. The insertion of a post-transcriptional regulatory element from the woodchuck hepatitis virus as well as of a Kozak consensus sequence led to a 7.5-fold increase in transgene expression. Human Epo production was further optimized by increasing the multiplicity of infection and by selecting high producer cells allowing to reach 200 IU hEpo/10E6 cells /day. These modified cells were encapsulated and implanted in vivo in the same conditions as previously described. All the mouse strains showed a sustained increase in their hematocrit and a high proportion of viable cells were observed after retrieval of the capsules. Finally, in the perspective of human application, a syngeneic model using encapsulated murine myoblasts transplanted in mice was realized to investigate the roles of both the host immune response and the cells metabolic requirements. Various loading densities and anti-inflammatory as well as immunosuppressive drugs were studied. The results showed that an immune process is responsible of cell death in capsules loaded at high cell density. A supporting matrix of PVA was shown to limit the cell density and to avoid early metabolic cell death, preventing therefore the immune reaction. This study has led to the development of encapsulated cells of human origin producing clinically relevant amounts of human EPO. This work resulted also to the optimization of cell encapsulation technical parameters allowing to begin a clinical application in end-stage renal failure patients. Résumé La thérapie cellulaire s'est imposée comme une stratégie de traitement potentiel pour diverses maladies. Si l'on considère leur morphologie et leur fonction, les cellules peuvent être transplantées dans le but de remplacer une perte tissulaire comme c'est le cas pour les transfusions sanguines ou les greffes de moelle osseuse ou de cellules pancréatiques. Avec le développement de la thérapie génique, les cellules sont également devenues des supports biologiques pour la production de molécules thérapeutiques. Cette stratégie représente le fondement de la thérapie génique ex vivo, caractérisée par le prélèvement de cellules d'un organisme, leur modification génétique et leur implantation dans le même individu ou dans un autre organisme. Le choix du type de cellule et la fonction qu'elle doit remplir pour un traitement spécifique dépend du tissu ou de la fonction biologique atteintes. Le but général de ce travail est de développer .une approche par thérapie génique ex vivo de sécrétion d'érythropoïétine (Epo) chez des patients souffrant d'anémie, prolongeant ainsi des travaux réalisés avec des cellules murines implantées chez des souris et des rats. Dans cette perpective, notre choix s'est porté sur des cellules humaines primaires allogéniques. En effet, contrairement aux cellules autologues, une caractérisation unique de cellules allogéniques peut déboucher sur de nombreuses applications. Par ailleurs, l'emploi de cellules allogéniques permet d'éviter les riques de zoonose que l'on peut rencontrer avec des cellules xénogéniques. Afin de protéger les cellules allogéniques soumises à une réaction immunitaire, leur confinement dans des macro-capsules cylindriques avant leur implantation permet d'éviter leur contact avec les cellules immunitaires de l'hôte, et de les retrouver sans difficulté en cas d'intolérance ou d'effet secondaire. Dans un premier temps, nous avons évalué la survie de différentes lignées cellulaires humaines primaires, une fois encapsulées et implantées dans le tissu sous-cutané de souris, soit immunocompétentes, soit immunodéprimées naturellement ou par l'intermédiaire d'un immunosuppresseur. Ce modèle in vivo correspond à des conditions xénogéniques et représente par conséquent un environnement de loin plus hostile pour les cellules qu'une transplantation allogénique. Une lignée fibroblastique issue du prépuce d'un jeune enfant, nommée DARC 3 .1, a montré une remarquable résistance avec un score de survie moyen le plus élevé parmi les lignées testées. Par la suite, nous nous sommes intéressés aux paramètres intervenant dans la réalisation du système d'implantation afin d'optimaliser les conditions pour une meilleure adaptation des cellules à ce nouvel environnement. En effet, en raison de l'apparition, après un mois in vivo, de calcifications au niveau de la matrice de polyvinyl alcohol (PVA) servant de support aux cellules encapsulées, différents paramètres ont été étudiés, tels que les procédures de fabrication, les lots de PVA ou encore les lignées cellulaires encapsulées, afin de mettre en évidence leur rôle respectif dans la survenue de ce processus. Nous avons montré que l'apparition des calcifications peut être totalement prévenue par l'utilisation d'eau pure au lieu de tampon phosphaté lors du rinçage des matrices de PVA. De plus, nous avons observe qu'un taux de lactate déshydrogénase cellulaire élevé était corrélé avec une diminution des dépôts de calcium au sein de la matrice en raison d'un micro-environnement plus acide inhibant la précipitation du calcium. Après sélection de la lignée cellulaire appropriée et de l'optimisation des conditions d'encapsulation, une modification génétique des fibroblastes DARC 3.1 a été réalisée par une approche rétrovirale, permettant l'insertion de l'ADN du gène de l'Epo dans le génome cellulaire. Diverses modifications, tant au niveau génétique qu'au niveau des conditions d'infection, ont été entreprises afin d'obtenir des taux de sécrétion d'Epo cliniquement appropriés. L'insertion dans la séquence d'ADN d'un élément de régulation post¬transcriptionnelle dérivé du virus de l'hépatite du rongeur (« woodchuck ») ainsi que d'une séquence consensus appelée « Kozak » ont abouti à une augmentation de sécrétion d'Epo 7.5 fois plus importante. De même, l'optimisation de la multiplicité d'infection et la sélection plus drastique des cellules hautement productrices ont permis finalement d'obtenir une sécrétion correspondant à 200 IU d'Epo/10E6 cells/jour. Ces cellules génétiquement modifiées ont été encapsulées et implantées in vivo dans les mêmes conditions que celles décrites plus haut. Toutes les souris transplantées ont montré une augmentation significative de leur hématocrite et une proportion importante de cellules présentait une survie conservée au moment de l'explantation des capsules. Finalement, dans la perspective d'une application humaine, un modèle syngénique a été proposé, basé sur l'implantation de myoblastes murins encapsulés dans des souris, afin d'investiguer les rôles respectifs de la réponse immunitaire du receveur et des besoins métaboliques cellulaires sur leur survie à long terme. Les cellules ont été encapsulées à différentes densités et les animaux transplantés se sont vus administrer des injections de molécules anti-inflammatoires ou immunosuppressives. Les résultats ont démontré qu'une réaction immunologique péri-capsulaire était à la base du rejet cellulaire dans le cas de capsules à haute densité cellulaire. Une matrice de PVA peut limiter cette densité et éviter une mort cellulaire précoce due à une insuffisance métabolique et par conséquent prévenir la réaction immunitaire. Ce travail a permis le développement de cellules encapsulées d'origine humaine sécrétant des taux d'Epo humaine adaptés à des traitements cliniques. De pair avec l'optimalisation des paramètres d'encapsulation, ces résultats ont abouti à l'initiation d'une application clinique destinée à des patients en insuffisance rénale terminale.

Construction and Quantitative Evaluation of a Dual Specific Promoter System for Monitoring the Expression Status of Stra8 and c-kit Genes.

Applications of genetic constructs with multiple promoters, which are fused with reporter genes and simultaneous monitoring of various events in cells, have gained special attention in recent years. Lentiviral vectors, with their distinctive characteristics, have been considered to monitor the developmental changes of cells in vitro. In this study, we constructed a novel lentiviral vector (FUM-M), containing two germ cell-specific promoters (Stra8 and c-kit), fused with ZsGreen and DsRed2 reporter genes, and evaluated its efficiency in different cells following treatments with retinoic acid and DMSO. Several cell lines (P19, GC-1 spg and HEK293T) were transduced with this vector, and functional capabilities of the promoters were verified by flow cytometry and quantitative RT-PCR. Our results indicate that FUM-M shows dynamic behavior in the presence and absence of extrinsic factors. A correlation was also observed between the function of promoters, present in the lentiviral construct and the endogenous level of the Stra8 and c-kit mRNAs in the cells. In conclusion, we recommend this strategy, which needs further optimization of the constructs, as a beneficial and practical way to screen chemical inducers involved in cellular differentiation toward germ-like cells.

Optimization of centrifugal separation of Ó-lactalbumin and ¿̐ư-lactoglobulin

Selostus: Ó-lactalbumiinin ja ¿̐ư-lactoglobuliinin sentrifugointierotuksen optimointi

Use of human MAR elements to improve retroviral vector production.

Retroviral vectors have many favorable properties for gene therapies, but their use remains limited by safety concerns and/or by relatively lower titers for some of the safer self-inactivating (SIN) derivatives. In this study, we evaluated whether increased production of SIN retroviral vectors can be achieved from the use of matrix attachment region (MAR) epigenetic regulators. Two MAR elements of human origin were found to increase and to stabilize the expression of the green fluorescent protein transgene in stably transfected HEK-293 packaging cells. Introduction of one of these MAR elements in retroviral vector-producing plasmids yielded higher expression of the viral vector RNA. Consistently, viral titers obtained from transient transfection of MAR-containing plasmids were increased up to sixfold as compared with the parental construct, when evaluated in different packaging cell systems and transfection conditions. Thus, use of MAR elements opens new perspectives for the efficient generation of gene therapy vectors.

An adenovirus-based fluorescent reporter vector to identify and isolate HIV-infected cells.

A procedure is described that allows the simple identification and sorting of live human cells that transcribe actively the HIV virus, based on the detection of GFP fluorescence in cells. Using adenoviral vectors for gene transfer, an expression cassette including the HIV-1 LTR driving the reporter gene GFP was introduced into cells that expressed stably either the Tat transcriptional activator, or an inactive mutant of Tat. Both northern and fluorescence-activated cell sorting (FACS) analysis indicate that cells containing the functional Tat protein presented levels of GFP mRNA and GFP fluorescence several orders of magnitude higher than control cells. Correspondingly, cells infected with HIV-1 showed similar enhanced reporter gene activation. HIV-1-infected cells of the lymphocytic line Jurkat were easily identified by fluorescence-activated cell sorting (FACS) as they displayed a much higher green fluorescence after transduction with the reporter adenoviral vector. This procedure could also be applied on primary human cells as blood monocyte-derived macrophages exposed to the adenoviral LTR-GFP reporter presented a much higher fluorescence when infected with HIV-1 compared with HIV-uninfected cells. The vector described has the advantages of labelling cells independently of their proliferation status and that analysis can be carried on intact cells which can be isolated subsequently by fluorescence-activated cell sorting (FACS) for further culture. This work suggests that adenoviral vectors carrying a virus-specific transcriptional control element controlling the expressions of a fluorescent protein will be useful in the identification and isolation of cells transcribing actively the viral template, and to be of use for drug screening and susceptibility assays.

Improvement of the quality factor of RF integrated inductors by layout optimization

A systematic method to improve the quality (Q) factor of RF integrated inductors is presented in this paper. The proposed method is based on the layout optimization to minimize the series resistance of the inductor coil, taking into account both ohmic losses, due to conduction currents, and magnetically induced losses, due to eddy currents. The technique is particularly useful when applied to inductors in which the fabrication process includes integration substrate removal. However, it is also applicable to inductors on low-loss substrates. The method optimizes the width of the metal strip for each turn of the inductor coil, leading to a variable strip-width layout. The optimization procedure has been successfully applied to the design of square spiral inductors in a silicon-based multichip-module technology, complemented with silicon micromachining postprocessing. The obtained experimental results corroborate the validity of the proposed method. A Q factor of about 17 have been obtained for a 35-nH inductor at 1.5 GHz, with Q values higher than 40 predicted for a 20-nH inductor working at 3.5 GHz. The latter is up to a 60% better than the best results for a single strip-width inductor working at the same frequency.

Analysis, modelling and classification of geospatial data using machine learning

The research considers the problem of spatial data classification using machine learning algorithms: probabilistic neural networks (PNN) and support vector machines (SVM). As a benchmark model simple k-nearest neighbor algorithm is considered. PNN is a neural network reformulation of well known nonparametric principles of probability density modeling using kernel density estimator and Bayesian optimal or maximum a posteriori decision rules. PNN is well suited to problems where not only predictions but also quantification of accuracy and integration of prior information are necessary. An important property of PNN is that they can be easily used in decision support systems dealing with problems of automatic classification. Support vector machine is an implementation of the principles of statistical learning theory for the classification tasks. Recently they were successfully applied for different environmental topics: classification of soil types and hydro-geological units, optimization of monitoring networks, susceptibility mapping of natural hazards. In the present paper both simulated and real data case studies (low and high dimensional) are considered. The main attention is paid to the detection and learning of spatial patterns by the algorithms applied.